QT Detailed 2 PDF

MP- 204
Vardhaman Mahaveer Open University, Kota
Quantitative Techniques
262
MP-204
263
MP-204

CONTENTS
Unit No.
Name of Unit
Page No.
BLOCK - I : BASIC MATHEMATICS

1.
Quantitative Techniques in Business
1-31
2.
Elementary Calculus and Matrix Algebra
32-64
BLOCK - II : FORECASTING
3.
Correlation and Simple Regression
65-85
4.
Business Forecasting and Time Series
86-108
5.
Index Numbers
109-133
BLOCK - III : PROBABILITY DISTRIBUTIONS

6.
Probability and Probability Distributions
134-155
7.
Decision Theory
155-175
BLOCK - IV : OPERATIONS RESEARCH

8.
Operations Research
176-193
9.
Linear Programming and Transportation Models
194-2198
10.
Networks
219-239
11.
Queuing and Game Theory
240-261
Unit-1 : Quantitative Techniques in Business

Unit Structure:
1.0
Objectives
1.1
Introduction
1.2
Meaning
1.3
Classification of Quantitative Techniques
1.4
Role of Quantitative Techniques
1.5
Limitations
1.6
Functions and Their Applications
1.7
Construction of Functions
1.8
Classification of Functions
1.9
Roots of Functions
1.10 Break-Even Analysis
1.11 Sequence and Series (Progression)
1.12 Arithmetic Progression
1.13 Geometric Progression
1.14 Applications of Arithmetic and Geometric Progressions
1.15. Summary
1.16 Key Words
1.17 Self Assessment Test
1.18 References
1.0
Objectives
After reading this unit, you should be able to understand the:
Meaning of Quantitative Techniques
Limitations of Quantitative Techniques
Functions and their applications
Construction and Classification of functions
Functions related to economics and some special functions
Roots of a function
Break Even Analysis
Sequence and Series
Arithmetic Progression, Sum of A.P. Series, Arithmetic Mean
Geometric Progression, Sum of G.P. Series, Geometric Mean
Applications of A.P. and G.P.

1
1.1
Introduction
Quantitative technique is a very powerful tool, by using this we can augment our production, maximize
profits, minimize costs, and production methods can be oriented for the accomplishment of certain pre
determined objectives. Quantitative techniques used to solve many of the problems that arise in a business
or industrial area. A large number of business problems, in the relatively recent past, have been given a
quantitative representation with considerable degree of success. All this has attracted the students, business
executives, public administrators alike towards the study of these techniques more and more in the present
times.
Scientific methods have been mans outstanding asset to pursue an ample number of activities. It is analyzed
that whenever some national crisis, emerges due to the impact of political, social, economic or cultural
factors the talents from all walks of life amalgamate together to overcome the situation and rectify the
problem. In this chapter we will see how the quantitative techniques had facilitated the organization in
solving complex problems on time with greater accuracy. The historical development will facilitate in
managerial decision-making & resource allocation, The methodology helps us in studying the scientific
methods with respect to phenomenon connected with human behaviour like formulating the problem, defining
decision variable and constraints, developing a suitable model, acquiring the input data, solving the model,
validating the model, implementing the results. The major advantage of mathematical model is that its facilitates
in taking decision faster and more accurately.
Managerial activities have become complex and it is necessary to make right decisions to avoid heavy
losses. Whether it is a manufacturing unit, or a service organization, the resources have to be utilized to its
maximum in an efficient manner. The future is clouded with uncertainty and fast changing, and decisionmaking a crucial activity cannot be made on a trial-and-error basis or by using a thumb rule approach.
In such situations, there is a greater need for applying scientific methods to decision-making to increase the
probability of coming up with good decisions. Quantitative Technique is a scientific approach to managerial
decision-making. The successful use of Quantitative Technique for management would help the organization
in solving complex problems on time, with greater accuracy and in the most economical way. Today, several
scientific management techniques are available to solve managerial problems and use of these techniques
helps managers become explicit about their objectives and provides additional information to select an
optimal decision. This study material is presented with variety of these techniques with real life problem
areas.
1.2
Meaning
Quantitative techniques are those statistical and operations research programming techniques which help in
the decision making process especially concerning business and industry. These techniques involve the
introduction of the element of quantities i.e., they involve the use of numbers, symbols and other mathematical
expressions. The quantitative techniques are essentially helpful supplement to judgement and intuition. These
techniques evaluate planning factors and alternative as and when they arise rather than prescribe courses of
action. As such, quantitative techniques may be defined as those techniques which provide the decision
maker with a systematic and powerful means of analysis and help, based on quantitative data, in exploring
policies for achieving pre determined goals. These techniques are particularly relevant to problems of
complex business enterprises.
1.3
We have many quantitative techniques in modern times. They can broadly be put under two categories:
(a) Statistical Techniques
(b) Programming Techniques
Statistical Techniques:
These techniques are used in conducting the statistical inquiry concerning a certain phenomenon. They are
including all the statistical methods beginning from the collection of data till the task of interpretation of the
collected data. More clearly, the methods of collection of statistical data, the technique of classification and
tabulation of the collected data, the calculation of various statistical measures such as mean, standard deviation,
coefficient of correlation etc, the techniques of analysis and interpretation and finally the task of deriving
inference and judging their reliability are some of the important statistical techniques.
Statistical Techniques also help in Correlation too. Correlation is a statistical technique that can show whether
and how strongly pairs of variables are related. For example, height and weight are related; taller people
tend to be heavier than shorter people. The relationship isnt perfect. People of the same height vary in
weight, and you can easily think of two people you know where the shorter one is heavier than the taller
one. Nonetheless, the average weight of people 55' is less than the average weight of people 56', and
their average weight is less than that of people 57', etc. Correlation can tell you just how much of the
variation in peoples weights is related to their heights.
Although this correlation is fairly obvious your data may contain unsuspected correlations. You may also
suspect there are correlations, but dont know which are the strongest. An intelligent correlation analysis
can lead to a greater understanding of your data.
There are several different correlation techniques. The Survey Systems optional Statistics Module includes
the most common type, called the Pearson or product-moment correlation. The module also includes a
variation on this type called partial correlation. The latter is useful when you want to look at the relationship
between two variables while removing the effect of one or two other variables.
Like all statistical techniques, correlation is only appropriate for certain kinds of data. Correlation works for
quantifiable data in which numbers are meaningful, usually quantities of some sort. It cannot be used for
purely categorical data, such as gender, brands purchased, or favourite colour.
Programming Techniques:
It can be defined as operational research or simply (O.R.) are the model building techniques used by
decision maker in modern times. They include wide variety of techniques such as linear programming,
theory of games, simulation, network analysis, queuing theory and many other similar techniques.
The following chart enlists the names of the important quantitative techniques:
(i)
Methods of collecting data
(i)
Linear Programming
(ii)
Classification and tabulation of

collected data
(ii)
Decision Theory
(iii)
Theory of Games
(iii)
Probability theory and Sampling
(iv)
Simulation:
Analysis
(iv)
Correlation and regression
(v)
Index number
(vi)
Time series Analysis
(vii)
Interpolation and Extrapolation
(viii)
Survey Techniques and
a. Monte Carlo Techniques

b. System Simulation
Methodology
(ix)
Ratio Analysis
(x)
Statistical quality control
(xi)
Analysis of Variance
(xii)
Statistical Inferences and
Waiting Line (queuing) Theory
(vi)
Inventory Planning
(vii)
Integrated Production Models
(viii)
Network Analysis/ PERT
(ix)
Others
a. Non- Linear Programming
b. Dynamic Programming
c. Search Theory
d. Integer Programming
Interpretation
(xiii)
(v)
e. Quadratic Programming
Theory of Attributes
f. Parametric Programming
g. The Theory of Replacement etc.
1.4
These techniques are especially increasing since World War II in the technology of business administration.
These techniques help in solving complex and intricate problems of business and industry. Quantitative
techniques for decision making are, in fact, examples of the use of scientific method of management. Their
role can be well understood under the following heads:
(i) Provide a tool for scientific analysis: These techniques provides executives with a more precise
description of the cause and effect relationship and risks underlying the business operations in measurable
terms and this eliminates the conventional intuitive and subjective basis on which managements used to
formulate their decisions decades ago. In fact, these techniques replace the intuitive and subjective
approach of decision making by an analytical and objective approach. The use of these techniques has
transformed the conventional techniques of operational and investment problems in business and industry.
Quantitative techniques thus encourage and enforce disciplined thinking about organisational problems.
(ii) Provide solution for various business problems: These techniques are being used in the field of
production, procurement, marketing, finance and allied fields. Problems like, how best can the managers
and executives allocate the available resources to various products so that in a given time the profits are
maximum or the cost is minimum? Is it possible for an industrial enterprise to arrange the time and
quantity of orders of its stock such that the overall profit with given resources is maximum? How far is
it within the competence of a business manager to determine the number of men and machines to be
employed and used in such a manner that neither remains idle and at the same time the customer or the
public has not to wait unduly long for service? And similar other problems can be solved with the help
of quantitative techniques.
4
(iii) Enable proper deployment of resources: It render valuable help in proper deployment of resources.
For example, PERT enables us to determine the earliest and the latest times for each of the events and
activities and thereby helps in identification of the critical path. All this helps in the deployment of the
resources from one activity to another to enable the project completion on time. This techniques, thus,
provides for determining the probability of completing an event or project itself by a specified date.
(iv) Helps in minimizing waiting and servicing costs: This theory helps the management men in minimizing
the total waiting and servicing costs. This technique also analyses the feasibility of adding facilities and
thereby helps the business people to take a correct and profitable decision.
(v) Assists in choosing an optimum strategy: Game theory is especially used to determine the optimum
strategy in a competitive situation and enables the businessmen to maximise profits or minimize losses
by adopting the optimum strategy.
(vi) They render great help in optimum resource allocation: Linear programming technique is used to
allocate scarce resources in an optimum manner in problem of scheduling, product mix and so on.
(vii) Enable the management to decide when to buy and how much to buy: The techniques of inventory
planning enables the management to decide when to buy and how much to buy.
(viii) They facilitate the process of decision making: Decision theory enables the businessmen to
select the best course of action when information is given to probabilistic form. Through decision tree
techniques executives judgement can systematically be brought into the analysis of the problems.
Simulation is an other important technique used to imitate an operation or process prior to actual
performance. The significance of simulation lies in the fact that it enables in finding out the effect of
alternative courses of action in situation involving uncertainty where mathematical formulation is not
possible. Even complex groups of variables can be handled through this technique.
(ix) Through various quantitative techniques management can know the reactions of the integrated
business systems: The Integrated Production Models techniques are used to minimise cost with
respect to work force, production and inventory. This technique is quite complex and is usually used by
companies having detailed information concerning their sales and costs statistics over a long period.
Besides, various other O.R. techniques also help in management people taking decisions concerning
various problems of business and industry. The techniques are designed to investigate how the integrated
business system would react to variations in its component elements and/or external factors.
1.5
Limitations
Quantitative techniques though are a great aid to management but still they cannot be substitute for decision
making. The choice of criterion as to what is actually best for the business enterprise is still that of an
executive who has to fall back upon his experience and judgement. This is so because of the several
limitations of quantitative techniques. Important limitations of these techniques are as given below:
(i) The inherent limitation concerning mathematical expressions: Quantitative techniques involve
the use of mathematical models, equations and similar other mathematical expressions. Assumptions are
always incorporated in the derivation of an equation and such an equation may be correctly used for the
solution of the business problems when the underlying assumptions and variables in the model are
present in the concerning problem. IF this caution is not given due care then there always remains the
possibility of wrong application of the quantitative techniques. Quite often the operations researchers
have been accused of having many solutions without being able to find problems that fit.
(ii) High costs are involved in the use of quantitative techniques: Quantitative techniques usually
prove very expensive. Services of specialised persons are invariably called for while using quantitative
techniques. Even in big business organisations we can expect that quantitative techniques will continue
to be of limited use simply because they are not in many cases worth their cost. As opposed to this a
typical manager, exercising intuition and judgement, may be able to make a decision very inexpensively.
Thus, the use of quantitative techniques is a costlier affair and this in fact constitutes a big and important
limitation of such techniques.
(iii) Quantitative techniques do not take into consideration the intangible factors i.e., non measurable
human factors: Quantitative techniques make no allowances for intangible factors such as skill, attitude,
vigour of the management people in taking decisions but in many instances success or failure hinges
upon the consideration of such non-measurable intangible factors. There cannot be any magic formula
for getting an answer to management problems; much depends upon proper managerial attitudes and
policies.
(iv) Quantitative techniques are just the tools of analysis and not the complete decision making
process: It should always be kept in mind that quantitative techniques, whatsoever it may be, alone
cannot make the final decision. They are just tools and simply suggest best alternatives but in final
analysis many business decisions will involve human element. Thus, quantitative analysis is at best a
supplement rather than, a substitute for management; subjective judgement is likely to remain a principal
approach to decision making.
1.6
Functions and Their Applications
Functions:
A function is a technical term used to symbolise relationship between variables. When one or more independent
variables are related to dependent variables then the dependent variable is said to be the function of independent
variables. A function explains the nature of relationship between dependent and independent variables. A
relationship may be a formula or a graph or a mathematical equation.
Examples of this type of relationship are price of a commodity and its demand, advertising expenses and
sales, income and expenditure, etc. This relationship is denoted by the concept of a function. If a dependent
variable Y is a function of an independent variable function is expressed as y = f(x) or y = F(x) and read as
Y is a function of x.
If the value of a set A belongs to x, generate another set B consists of values of y, then the function is
expressed as f: A!B and read as f is a function from A into B or A maps into B.
Here set A is called the domain of the function and set B is called the range of the function. For any
orbitrarity value a belongs to x, gives a unique value f (a) corresponding to y.
Generally for n independent variables x1, x2, x3 xn, and y is a dependent variable the function
can be given as y = f (x1, x2,x3 , xn )
Some examples of functions are given by
y = x - 15
0<x<1
y = x2 + 6x - 2
1< x < 2
In (i), the domain of x is the closed interval [0, 1] and the range of y is [-15, -14]. In the (ii), the domain of
x is the interval [1, 2] and the range of y is [5, 14].
6
1.7
Construction of Functions
Many times in business, we commonly talk about profit functions, loss functions, cost functions and revenue
functions, production function, demand and supply function, consumption function, into function, onto function,
polynomial function, Absolute value function, Step function, Inverse function, Rational and Irrational function,
Algebraic function. The functions are usually set up following the definition and calculation of the functional
values. Now we will take some few examples to illustrate the method of constructing such functions.
Example 1:A factory has 100 items on hand for shipment to a destination at the cost of Rs 1 a piece
to meet a certain demand d. In case the demand d overshoots the supply. It is necessary to meet the
unsatisfied demand by purchases on the local market at Rs 2 a piece. Construct the cost function if
x is the number shipped from the factory.
Solution: Suppose C(x) denotes the cost function.
If d x,
C(x) = x + 2(d-x)
If d < x,
C(x) = x
It can be combined into a single representation by writing

C(x) = x + 2 max (0, d-x)
This definition of C(x) is seen to be equivalent to the earlier one by considering separately the cases
d x and d < x.
Example 2: If f(x) = 2x + 1, find the range if domain is {-1,2,3} and hence find the function.
Solution: Here f(x) = 2x + 1
=
=
f(-1) = -1 2 + 1 = -1,
f(2) = 2 2 + 1 = 5,
= f(3) = 3 2 + 1 = 7
So,
Range = {-1, 5, 7}
And
f = {(-1, -1), (2, 5), (3, 7)}
1.8
Classification of Functions
We have already learnt that for a function f : A B, f associates all elements of set A to set B and each
element of set A is associated to a unique element of a set B. Thus, we may associate different element of set
B or we may associate more than one element of set A to same element of set B (but same element of set A
cannot be associated with more than one element of set B). Also, all elements of set B may or may not have
their pre-images in A.
1.8.1 Functions Related To Economics:
In the case of functions in economics, the variables are hypothetical quantities and not actual observable
quantities as in physical science.
The range and the domain of economics functions are made up of nonnegative quantities so that the graphs
of these functions are in first quadrant only. In other words, for the purpose of economic analysis, only that
part of a curve is relevant which lies in the first quadrant.
1. Demand Function:
As we know that the quantity demanded of a particular commodity by the buyers in the market is depending
on the price. As the prices increases, the demand is decreases shown as figure. If q is the quantity of a
commodity demanded and p is the price then the demand function is given by
q = f (p) shows q depends on p price
p = g (q) shows p depends on q
for example: q = a - bp or q = (10/p)
q = - 10p2 + p + 5
or
P
Price
Demand
X
Demand Function
2. Supply function:
Price of any particular commodity in the market depends on the quantity of supply. As the quantity of supply
increases the price is also increases. If x is the quantity of supply and p is its price then the supply function
is given by x = f (p) or p = f (x).
P
Price
Supply
X
Supply Function
3. Cost Function:
3.1. Total Cost Function
3. 2. Average Cost Function
3.1 Total cost function: The total cost (C) which is equal to sum of fixed cost and variable cost, of
production in a firm, is depending on the quantity produced of a particular commodity. As the production
increases the total cost also increases. If x is the quantity produced at total cost T then the total cost
function is given by T = f (x).
3.2 Average Cost Function: Average cost function or cost per unit is obtained by dividing the total cost
function by the quantity produced. Average cost function A = (T / x) = (f (x) / x).
4. Total Revenue Function:
The total revenue in a particular commodity of a firm is depending on the number units sold. If p is the price
of any commodity and q is the quantity sold then the total revenue function is given by R = p.q, and the
average revenue function or revenue per unit is obtained by dividing the total revenue function by the units
of quantity sold, it is equal to price (p) of the commodity.
AR = (R / q) = (Pq / q) = p = price
5. Profit Function:
Profit on a particular commodity of a firm is equal to the difference or total revenue and the total cost the
profit function is given by z = R T.
6. Production Function:
The output product is depending in the various inputs like capital, labour, raw material etc. If q is the output
quantity of a particular commodity and x, y and z are the input variables in a firm then the production
function as sin y.
q = f (x, y, z)
For example: If labour (1) and capital (K) are inputs variables then the output is given by Q = f (L, K) =
AL - K, + = 1 where A, and are constants is known as Cobb-Douglas production function.
7. Consumption Function:
The total consumption function of a firm is depending on the income as income increases the consumption
expenditure also increases. If C is the consumption and I is the income then the consumption function given
by
C = f (5) = a + bI where a, b are constants.
Example1: A company sells x tins of chicken each day at Rs. 80 per tin. the cost of production and
selling price of these tins is Rs. 50 per tin plus a fixed daily overhead cost of Rs. 18,000. Determine
the profit function. What is the profit if 2,000 tins are produced and sold a day? Find out the number
of tins produced in a day with no profit and no loss.
Solution: a). Per day x tins of chicken are produced with cost Rs. 80 per tin the revenue received per day
is Revenue function
R (x) = 80. x
The cost of production per day cost function is c(x) = 1,800 + 50x
If P (x) is the profit function is given by
P (x) = R(x) - c(x)
= 80x - (50x +1,800) = 30x - 1,800
If 2,000 tins are produced and sold in a day then the profit is given by
P (2,000) = 30 2,000 - 18,000
= 60,000 - 18,000 = Rs. 42,000
b). for no profit and no loss then profit function is zero
P (x) = 0 = 30x - 18,000 = 0
9
30x = 18,000x = (18,000 / 30) = 600

600 tins per day.
1.8.2 Some Special Functions:
In this case, we have some special functions which can be used in economics as well as computer science
area too. These functions are defined as follows:
1. Polynomial function:
The general form of the function
f (x) or y = f (x1, x2, x3,.., xn ) = an xn + an-1 xn-1 + an-2 xn-2 + + a0
(i)
Where ais are real numbers, a1 0, and n is positive integer is called a polynomial of degree n. The real
constants an, an-1, an-2, a0 are called coefficients. The exponent n is called the degree of the polynomial. For
n = 1, the equation (a) can be written as
y = a1 x + a0
(a 0)
This is usually written as

y = a + bx
(b 0)
Where a = a0 and b = a1. Such Polynomial functions are called linear function
(i. e. degree 1) and has the domain {x: x R}.
(b) Now for n = 2, the equation (i) can be written as:
y = a2 x2 + a1 x1 +a0
(a2 0)
or
y = ax2 + bx + c, (a 0)
Where a = a2, b = a1 and c = a0. Such polynomial functions are called second degree functions or quadratic
functions.
The x -intercept of a quadratic function is very important. The x- intercept is that point (or points) at which
the graphs crosses the x-axis; that is, the point at which y = 0. If such points exist, they satisfy the equation:
ax2 + bx + c = 0.
The graph of the linear function is the straight line y = a + bx which cuts the x-axis at a distance -b/a units
from the origin and y-axis at a distance of a units from the origin as shown.
Y
A(0,b)
y = a + bx
B(- b / a,0)
Inverse Mapping
Functions like f(x) = x3 + 2 x

g(x) = 2 + x - x4
are polynomial function but 2x + x 2/3 is not a polynomial function.
10
Example 1: The demand for certain item is given by q = 150 - 3p Where q denotes the amount
demanded and p the price per unit. It costs Rs 4 to produce each unit. What is the profit function
of the firm for this item?
Assume Y = total profit
R = total revenue
C = total cost
Since
R = pq, C = 4q,
Profit = Total Revenue Total cost

Y= R - C
= pq - 4q
= q (p - 4)
= (150 3p) (p 4)
[Substituting q = (150 3p) given]
= -3p2 + 162p -600 which is a quadratic function in p.

2. Absolute Value Function:
The functional relationship of the form f(x) or y = |x| is known as an absolute value function, where x
represents the absolute value of x and is defined as given below, and the graph of this function is also shown
y = |x| =
x, for x = 0
-x,
x<0
0,
x=0
3. Step Function:
If a function is defined on a closed interval [a, b] and assume a constant value in the interior of each subinterval say [a, x1], [x2, x3], . , [xn, b] of [a, b] where a < x1 < x2 < < xn < b, then such function
is called a step function. Symbolically it may be expressed as:
y or f (x) = ki for all values of x in the ith sub-interval.
The graph of this function is given below:
Y
y3
y2
y1
O
X
Step Function
11
4. Convex Set and Convex Function:

A set S of points in the two-dimensional plane is said to be convex if for any two points (x1, y1) and (x2, y2)
in the set the line segment joining these points is also in the set.
Mathematically, this definition implies (x1, y1) and (x2, y2) are two different points in S, Then the point whose
coordinates given by
{ x1 + (1- ) x2; y1 + (1 - ) y2}; 0 1
must also be in the set S.
If = 0, then we get the coordinates (x2, y2) n the given point. But, if = , the corresponding point on the
line segment is [(x1 + x2)/2, (y1 + y2)/2]. This point lies in the centre of the line segment joining two points(x1,
y1) and (x2, y2).
The examples of convex set are circle and triangle. The examples are given below for convex and non
convex sets.
Convex function: A function f(x) defined over a convex set S is said to be convex function if for any two
distinct points x1 and x2 lying in S for any 0 1, f { x1 + (1- ) x2} f (x1) + (1- ) f (x2).
5. Inverse Function:
If variable x and y are interdependent such that (i) y is the function of x; y = f(x)
(ii) If x is the function of y, x = g (x), then f is termed as the inverse of g and vice-versa. For example, if y=
x2 + 2x + 6, then the inverse function is: x = -1
y 6
These functions are related to as follows:

y = f (x) = f [g (y)] or fog (y)
and
x = g (y) = g [f (x)] or gof (x)
Consequently fog = gof. Functions fog and gof are termed as composite functions.
6. Rational functions:
A rational function is defined as the quotient of two polynomial functions and is of the form:
an xn + an-1xn-1 + + a0
y=
bm x
m-1
+ bm-1x
+ .+ b0
= P(x)
, Q(x) 0
Q(x)
Where P (x) is a polynomial function of degree n and Q (x) is a non zero polynomial function of degree m.
12
An expression which involves root extraction on terms involving x is called irrational function. The function
such as
x,
2 x 2 7 x 6 are examples of irrational functions.
7. Algebraic Function:
A function consisting a finite number of terms involving powers and roots of the variable x and the four basic
mathematical operations (addition, subtraction, multiplication and division) is called an algebraic function. In
general, it can be expressed as
yn + A1yn-1 + . + An = 0 where A1, A2, .. , An are rational functions of x.
There are two categories of algebraic functions namely: explicit and implicit algebraic functions. For example,
y=
2
2
x 2x 3 is and enplicat algebraic function, where as xy + xy + x = 0 is an implicit function.
8. Transcendental Function:
All functions which are not algebraic are called transcendental functions. These functions include:
1. Trigonometric Functions:
The trigonometric functions of an angle ( be any real number) are given by:
sin = sin c,
cos = cos c, tan = tan c
cosec = cosec c, sec = sec c , cot = cot c
where denote the angle whose radian measure is in .
The sin and cosec are said to be co - functions, as are the cos and sec and tan and cot.
The trigonometry functions are defined similarly for negative angles. An angle may be measured in degrees
or radians. However, in calculus and its applications to business and economics radian measure is usually
more convenient.
The trigonometric functions are very useful in the study of business cycles, seasonal or other cyclic variations
are described by sine or cosine functions.
2. Exponential Functions:
A function having a constant base and a variable exponent is called an exponent function, such as
(i)
(iii)
y = ax, a 1, a > 0
y = k abx a 1, a > 0
(ii)
(iv)
y = k ax a 1, a > 0
y = k ex
where a, b, e, and k are constants and x is exponent.

In calculus and its applications to business and economics, such functions are useful for describing sharp
increase and decrease in the value of dependent variable. For example the graph of exponential function y
= kax indicates rise to the right in the value of y for a > 1 and k > 0 whereas indicates fall to the left for a <
1 and k >0, as shown in the fig.
Y
y = kax
(a> 1, k>0)
(a)
13
The rules governing the exponents are as under:

(i) a x1 . ax2 = a x1+ x2
(ii) a x1 /ax2 = a x1- x2
(iii) (a x1) x2 = a x1 x2
(iv) (a.b) x1 = a x1. b x1
(v) (a /b) x1 = a x1. b- x1
(vi) a0 = 1.
3. Logarithmic Functions:
A logarithmic function is expressed as loga x where a > 0, a 1 is the base. It is read as y is the log to the
base a of x. This relationship may also be expressed by the equation x = ay. It is an exponential function.
Thus, logarithmic and exponential functions are inverse functions, i.e. if x is an exponential function of y, then
y is a logarithmic function of x.
Although the base of logarithm can be any positive number other than 1, but most widely used bases are
either 10 ( common or Briggsian logarithms) or e = 2.718 (natural or Naperian logarithms). By convention,
log x denotes the common logarithm of x in x denotes natural logarithm of x. If any other base is meant, it is
specified.
Some important properties of the logarithms are as follows. If x and y are positive real number, then
(i) loga xy = loga x + loga y
(ii) loga (xy) = loga x - loga y
(iv) loga x1/n =(1 / n) loga x
(v) loga x = loga b logb x
(iii) loga xn = n loga x
(vi) Logarithm of zero and negative number is not defined.

Since exponential function x = ay and logarithmic function y = loga x are universe functions, therefore graph of
these curves for a particular value of a can be obtained from the graph of each other by taking reflection
about the line y = x, as shown in the fig
Incommensurable Power Functions: A function having a variable base and a constant exponent is called and
incommensurable power function, such as y = x
or x 3/2 or xa etc.
9. Even and Odd Functions:

If a function not change its sign of its independent variable changed, then it is said to be an even function, i.e.
f (-x) = f(x). The examples of even functions are x6, cos x, etc. It follows that the graph of an even function
is symmetrical about the y - axis.
On the other hand, f(x) is said to be odd if f(-x) = -f(x). The examples of odd functions are: x7, 10. Periodic
Function: If f(x + T) = f(x), where T is a real number, then f(x) is called a periodic function. The real number
T is called a period of f(x).
The least positive period of a periodic function is called the principal period of that function. Since for all real
numbers x, sin(2 + x) = sin x and cos (2 + x ) = cos x, therefore the function sin x and cos x are periodic
functions with period 2. Other than this 4, 2, 6 etc. are also periods of sin x and cos x.
14
Sine Function
1.9
Roots of Functions
The value (or values) of x at which the given function f(x) becomes equal to zero are called zeros of the
function f(x). The zeros of the function are also called the roots of the given function f(x).
For the linear function y = ax + b the roots are given by ax + b = 0, i. e. x = - b/a.
Thus, if x = - b/a, then substituting in the given equation, ax + b = 0, the left hand side of its becomes equal
to its right hand side. For the quadratic function y = ax2 + bx + c, a 0 the quadratic equation ax2 + bx +
c = 0, a 0 has to be solved to find the roots of the function y. The general value of x which satisfy the given
quadratic equation is given by
This shows that, in general, there are two values of x (also called roots) for which
ax2 + bx + c become zero. The two values are
Note: The numbers of roots of the given function are always equal to the highest power of the independent
variable.
1.10 Break-Even Analysis

Break Analysis is often referred as cost volume profit analysis. This is because the purpose of the
analysis is to study the relationship among cost, volume and profit in order to determine the point of breakeven volume. The major components of the break even model are:
(i) Total Revenue: It consists of selling price (p) multiplied by the quantity sold (q) i.e., R = p.q
(ii) Fixed Cost: The cost which remains constant (in rupees) regardless of the volume of production. Let it
be Rs. k.
(iii) Variable Costs: The costs which vary with the quantity produced or sales. It is calculated by multiplying
cost/unit of the output with the volume of production, i.e., V = v.q, where v is the variable cost/unit of
output.
15
The break - even analysis seeks to determine the output at which the producers will break-even, i.e. the
level of output where his costs will exactly meet his revenue.
The total profit associated with an output of q units is given by
Profit (P) = Total revenue Total cost
= (Price) (Quantity) {Fixed cost + (Variable cost) (Quantity)}
= p . q ( k + v . q)
By definition, break- even analysis determines the optimum value of q for which profit P equals
zero, i.e. Total revenue = Total cost
or
p . q ( k + v . q) = 0
q* (optimum)
=
k
pv
Fixed Cost
Selling Price Variable Cost
Solved Examples
Example 1. A firm produces an item whose production cost function is C = 80 + 4x, where x is the
number of items produced. If entire stock is sold at the rate of Rs. 8 then determine the revenue
function. Also obtain the break even point.
Solution. The revenue function is given by R = 8x. Also given that , C = 80 + 4x. Therefore
Profit, P = R C = 8x (80 + 4x) = 4x 80
The break- even points occur when R C = 0 or R = C, i.e. 8x = 80 + 4x or x = 20 (units).
Example 1. A company producing dry cells introduces production bonus for its employees which
increases the cost of production. The daily cost of production C(x) for x number of cells is Rs. (3.5x
+ 12,000).
(a) If each cell is sold for Rs. 6 , determine the number of cells that should be produced to ensure no
loss.
(b) If the selling price is increased by 50 paise, what would be the break even point?
(c) If at least 6000 cells can be sold daily, what price the company should charge per cell to guarantee
no loss?
Solution. Let R(x) be the revenue due to the sales of x number of cells.
(a) Given that, cost of each cell is Rs. 6. Then R(x) = 6x. For no loss, we must have
R(x) = C(x) or 6x = 3.5x + 12,000 or 12,000/2.5 = 4,800 cells.
(b) Increased selling price is, Rs.(6 + 0.50) = Rs. 6.5. Thus, R(x) = 6.5x. Now, for break- even point, we
must have
R(x) = C(x) or 6.5x = 3.5x + 12,000 or x = 12,000/3 = 4000 cells.
(c) Let p be the unit selling price. Then revenue from the sale of 6000 cells will be, R(p) = 6000 p.
Thus, for no loss we must have
R(p) = C(p) or 6000p = 3.5 6000 + 12,000 or p = 33,000/6,000 = Rs. 5.5.
16
1.11 Sequence and Series (Progression)

1.11.1 Sequences
A sequence is a function, whose domain is the set N of natural numbers or rule set of N. We may also define
a sequence as an arrangement of numbers according to a certain fixed rule. The different numbers in
sequences are called its terms and are generally denoted by a1, a2, a3, .. or T1, T2,T3.....etc. Some examples of
sequences are:
1, 3, 5, 7, 9,
(i)
1, 3, 9, 27, ..
(ii)
In (i), the terms are arranged such that the difference between two consecutive terms is the same i.e. , 2.
In (ii), the terms are arranged in such a way that the ratio of two consecutive terms is the same i.e., 3.
Note. It is obvious that if the rule (which the terms of a sequence obey) is known, then all or any
term can be found.
General Term
The nth term of a sequence is called the general term and is denoted by an or Tn .
Finite and Infinite Sequences
Finite Sequence: A sequence containing finite number of terms is called a finite sequence.
For example, 2, 4, 6, 8, , 40
1, 3, 9, , 243
24, 21, 18, upto 6 terms.
Infinite Sequence: A sequence is called infinite if it is not a finite sequence or in other words, a
sequence having infinite terms is called an infinite sequence.
For example, 1, 3, 5, 7, , to infinity
3, 33,333,
Note. If the general term (=Tn ) of a sequence is given, we can find T1, T2,T3..... by putting n = 1, 2,
3, . in Tn .
1.11.2 Series
The terms of a sequence connected by +ve or - ve sign from a series.
For example,
1+ 3+ 5+ 7+
The different terms of the series are denoted by or T1, T2,T3.....

Solved Examples
Example 1. Write the first three terms of the sequences whose nth terms are
(i) 3n - 5
(ii) 2n
(iii) n2
Solution. (i) Here Tn = 3n - 5

Putting n = 1, 2, 3 in Tn , we get
T1 = 3.1 - 5 = - 2
T2 = 3.2 - 5 = 1
T3 = 3.3 - 5 = 4
(ii) Here Tn = 2n
17
T1 = 21 = 2
T2 = 22 = 4
T3 = 23 = 8
(iii) Here Tn = n2
T1 = 12 = 1
T2 = 22 = 4
T3 = 32 = 9
Example2. If Tn = n2, when n is odd and Tn = 2n, when n is even, then write down the first three
terms.
Solution. Here Tn = n2, when n is odd
Putting n = 1, 3, 5 , we get
T1 = 12 = 1
T3 = 32 = 9
T5 = 52 = 25
Also, Tn = 2n, when n is even

Putting n = 2, 4, 6 , we get
T2 = 2.2 = 4
T4 = 2.4 =8
T6 = 2.6 =12
Hence, the first 6 terms of the sequence are 1, 4, 9, 8, 25, 12.

Example3. If Tn = (n - 1) (n - 2) (n - 3); show that the first three terms of the sequence are zero, but
the rest of the terms are positive.
Solution. Here Tn = (n - 1) (n - 2) (n - 3)
Putting n = 1, 2, 3 in Tn successively, we get
T1 = (1 - 1) (1 - 2) (1 - 3) = 0, T2 = (2 - 1) (2 - 2) (2 - 3) = 0,T3 = (3 - 1) (3 - 2) (3 - 3) = 0
For n>3, each of the terms (n - 1) (n - 2) and (n - 3) is positive, and hence their product (n - 1) (n - 2) (n
- 3) is also positive.
Hence, the first three terms are zero, but the rest of the terms are positive.
1.12 Arithmetic Progression

An arithmetic progression is a sequence whose terms increase or decrease by a constant number called the
common difference. A series in arithmetic progression thus becomes an additive series in which the common
difference can be found by subtracting each term from its preceding one. Thus,
(i)
The sequence 1, 5, 9, 13, 17, 21, 25 is an infinite arithmetic progression of seven terms, the
first term is 1 and the common difference is 4. Similarly,
(ii) The sequence 12, 5,- 2, - 9, is an infinite arithmetic progression of four terms , the first term is
12 and the common difference is 7.
The corresponding arithmetic series are
(i)
1 + 5 + 9 + 13 + 17 + 21 + 25 +
(ii) 11 + 17 + 23 + 29 + .
Thus if the first term and common difference are known, the A.P. is completely known,
The arithmetic progression a, (a + d), (a + 2d), (a + 3d)
Whose first term is a and the common difference is d, is designated as the standard form of the arithmetic
progression.
18
The corresponding arithmetic series a + (a + d) + (a + 2d) + (a + 3d) +

is designated as the standard form of an arithmetic series. The abbreviation A. P. for arithmetic
progression is commonly used. In other words, a sequence in which Tn - Tn-1 = constant difference (say d), is
called an A. P.
This difference is called the common difference and is denoted by d.
Note. 1. In general, a, a + d, a + 2d represents an A.P. whose first term is T1 and common difference is
d.
2. nth term (T1) of an A. P. is called its general term.
nth term of an A. P.
Let the A. P. be an a, (a + d), (a + 2d), (a + 3d), whose first term = a and common difference = d
Then,
T1 = a = a + (1 - 1) d = a + 0. d
T2 = a + d = a + (2 - 1) d
T3 = a + 2d = a + (3 - 1) d = and so on. So, nth term of an A. P. is
Tn = a + (n - 1)d
Properties of an A.P.
If each of the term of an A. P. is
(i) increased by (ii) decreased by
resulting sequence is still in A.P.
(iii) multiplied by
(iv) divided by a constant quantity, then the
Let the A.P. be a, (a + d), (a + 2d), .. and k is a constant quantity then, if we add, subtract, multiplied or
divide the series by k then it will still remain an A.P.
Example 1. Find the 12th term of the A. P. 3, 7, 44, 15
Solution. Here a = 3, d = 7 - 3 = 4
So, Tn = a + (n - 1) d
T12 = 3 + (12 - 1)4
= 3 + 44 = 47
Example 2. Which term of series 12 + 9 + 6 + is equal to (i) - 30, (ii) - 100?
Solution. (i) The series in an A.P. with first term 12 and common difference -3.
Tn = a + (n - 1) d = 12 + (n - 1) (- 3) = 15 - 3n
Suppose the nth term is - 30 then 15 - 3n = - 30 n = 15
So, - 30 is the 15th term.
(ii) Now suppose that Tn = - 100, 15 - 3n = - 100 n = 115
3
Which is impossible, because n must be a whole number. Hence there exists no term in the series
which is equal to - 100.
19
Sum of the first n terms of the general A. P.:

The sum of a series in A. P. is an important quantity which yields many other related results. We denote the
sum of n terms by Sn and the first and last terms of the sequence by a and l respectively. Let Sn denote the
sum of n terms, then
Sn = a + (a + d) + (a + 2d) ++ (l - 2d) + (l - d) + l
. (1)
Writing the above series in reverse order, we have

Sn = l + (l - d) + (l - 2d) ++ (a + 2d) + (a + d) + a
(2)
Adding (1) and (2), we get

2Sn = (a + l) + (a + l) + (a + l) + up to n terms = n (a + l)
Sn
l = a + (n - 1) d
But
Sn
n(a 1)
2
n[a a (n 1)d ]
2
Note: Generally, formula (I) is used when the last term is given and (II) is used when
Common difference is given.
Example 1. Find the sum of the series: 72 + 70 + 68 + + 40
Solution. The given series is an A. P. with a = 72 and d = -2. Let n be the number of terms. Then
Tn = 40.
a + (n - 1) d = 40 72 + (n - 1) (-2) = 40, i. e., 2n = 34 or n = 17.
Now
n(a 1)
2
Example 2. The first and the last term of an A. P are respectively - 4 and 146, and the sum of the
A.P. = 7171. Find the number of terms of the A.P. and also its common difference.
Sn
Solution. We have a = - 4 and l = 146 and Sn = 7171

Let n be the number of terms of the A.P.
Sn
n(a 1)
n(4 146)
7171 or , Sn
7171
2
2
n = 101, Also 146 = (- 4 ) + (101 - 1) d 100d = 146 + 4 = 150

d = 1.5. Hence n = 101 and d = 1.5
Example 3. Find the sum of all numbers between 100 and 1000. Which are divisible by 11?
Solution.
Numbers between 100 and 1000 divisible by 11 are 110, 121, 132 990, which
clearly form an A.P., where first term, a = 110; the common difference d = 11, and the nth term Tn
= 990. Then, we have Tn = a + (n - 1) d
990 = 110 (n - 1) 11, i.e., n = 81.
So the sum of these numbers is given by
S n = n (a + l)
2
Sn = 81(110 +990) = 44550

2
20
Example 4. Find the series whose sum to n terms is 5 n + 3n.

2
Solution. Given that Sn = 5 n + 3n. Pitting n = 1, and 2, we get

S1 = 5 (1)2 + 3(1) = 8 and S2 = 5 (2)2 + 3(2) = 26
Now first term, a = S 1 =8 , second term S2 - S1 =26 - 8 = 18. Then the common difference, d =
second term - first term = 18 - 8 = 10.Hence, the required A.P. is: 8, 18, 28, 38
Some Important Formulae
(a) 1 + 2 + 3 + + n =
n (n + 1)
2
(b) 1 + 2 + 3 ++ n = n(n + 1)(2n +1)
2
(c) 1 + 2 + 3 ++ n =
2
2
n (n + 1)
2
Arithmetic Mean (A.M.)

If a, A, b are in A.P., then A is the A.M. between a and b, i.e., if three quantities are in A.P., then the
middle one is defined as A.M. between the other two.
Arithmetic Mean Between a and b
Let A be the A.M. between a and b.
So, a, A, b are in A.P.
Thus,
A - a = b - A, i.e.,
Similarly,
A=
2A = a + b
a+b
2
To insert n A.Ms between two quantities a and b

Let A1, A2, A3, , An be the n A.Ms between a and b.
So, a, A1, A2, A3 An, b are in A.P.
In this A.P., the first term is a, the last te rm is b and the num ber of term s is (n + 2) i.e., b is the
(n + 2)th term= a + (n + 2 - 1)d = a + (n + 1)d
d=
ba
A1 = a + d = a + b - a
n+1
n+1
A 2 = a + 2d = a + 2. b - a
n +1
A n = a + nd = a + n. b - a ,
n+1
Whic h are the required a rithm etic means betwe en a and b.
21
Example 1. The first and the eight terms of an A.P. series are 2 and 23 respectively. Find the tenth
term of the series.
Solution. Let common difference = d
T8 = 2 + ( 8 - 1)d 23 = 2 + 7d d = 3, Tenth term T10 = 2 + (10 - 1)3 = 29
Example 2. The pth, qth, and rth terms of an A.P. series are a, b, and c respectively. Prove that, (q r)a + (r - p)b + (p - q)c = 0
Solution. Let the first term = a, and the common difference = d
According to the problem we have.
a = a1 + (p - 1) d
(1)
b = a1 + (q - 1) d
(2)
c = a1 + (r - 1) d
(3)
Now, (2) - (3) b - c = (q - r) d
(4)
(3) - (1) c -a = (r - p) d
(5)
(1) - (2) a -b = (p - q) d
(6)
Now, Multiplying (4), (5), and (6) by a, b, and c respectively and then adding we get,
(b - c)a + (c - a)b + (a - b)c = {(q - r)a + (r - p)b + (p - q)c}d
0 = {(q - r) a + (r - p) b + (p - q) c} d
(q - r) a + (r - p) b + (p - q) c = 0
Note. The relation to be shown contains (q - r), (r - p), and (p - q). If we subtract (3) from (2), (1) from (3),
and (2) from (1) then and only then we will get (q - r), (r - p), and (p - q).
Example 3. How many terms of this series 27 + 24 + 21 + 18 + will add up to 126?
Solution. Let the first n terms of this series add up to 126.
Sn= 126, a= 27, d = - 3, n =?
n[2a (n 1) d ]
2
n[54 ( n 1)d ]
Sn
2
n[54 (n 1)( 3)]
126
2
Sn
252 = n [54-3n+3]
252= n [57-3n]
252 = 57n-3n2
n2 19n +84 = 0
n2 7n 12n + 84 = 0
(n - 7) (n 12) = 0
n = 7 or n = 12
i.e., the sum of the first 7 terms is 126 and the sum of first 12 terms is also 126.
22
Example 4. If the third and the sixth terms of an A.P. are 7 and 13 respectively. Find the first term
and the common difference.
Solution. Let the first term of the A.P. series is a and its common difference is d.
Now, Tn = a + (n - 1) d T3 = a + (3 - 1) d 7 = a + 2d
Tn = a + (n - 1) d T6 = a + (6- 1) d 13 = a + 5d
(2) - (1)
3d = 6 d = 2
Now, putting the value of d in (1) we get a = 3, the 1st term = 3 and common difference = 2.
Example 5. The sum of three numbers in A.P. is 33 and their product is 1287, Find the numbers.
Solution. Let the three numbers be a - d, a, a + d. Clearly, these three numbers are in A.P.
According to the question a - d + a + a + d = 33
3a = 33 a = 11
Again, (a -d). a. (a + d) = 1287 a (a2 - d2) = 1287 11(112 - d2) = 1287 121 - d2 = 117
d2 = 4 d = 2
When d = 2, then the three numbers will be: 11 - 2, 11, 11 + 2 i.e., 9, 11, 13
Again, when d = - 2, then the three numbers will be: 11 - (- 2), 11, 11 + (- 2) i.e., 13, 11, 9.
1.13 Geometric Progression

A geometric progression is a sequence whose terms increase or decrease by a constant ratio called the
common ratio. A series in geometric progression thus is a multiplicative series whose common ratio can be
found by dividing any term by its preceding term. Thus
(i) The sequence 1, 2, 4, 8, 16, 32 is an infinite geometric progression, the first term is 1 and the
common ratio is 2. Similarly,
(ii) The sequence 5, -10, 20, -40, 80 is a geometric progression, the first term is 5 and the common
ratio is -2.
The corresponding geometric series are:
1 + 2 + 4 + 8 + 16 + 32 +
5 - 10 + 20 - 40 + 80 -
The geometric progression is, therefore, in the form: a, ar, ar2, ar3, ar4
Whose first term is a and the common ration is r, and is designated as the standard form of a progression.
The corresponding geometric series is a + ar + ar2 + ar3 + ar4 + and is designated as the standard form
of geometric series. The abbreviation commonly used for geometric progression is G.P.
In other words, a sequence of non-zero numbers a1, a2, a3 an is called a geometric

a k 1
progression if
r (constant), for k 1 and k N.
ak
nth term of a G.P. is called its General Term.
General Term of a G.P.
Let us consider a G.P. with first non-zero term a and common ratio r. Then,
T1 = a = ar1 1
T2 = ar = ar2 1
T3 = a r2 = ar3 1 Tn = ar n - 1
23
Thus the nth term of a G.P. is given by Tn = ar n - 1

Remark. If the nth term is the last term and is denoted by l, then l = ar n - 1.
Solved Examples
Example 1. Which term of the G.P.? 5, 1, 0.2 is 0.0016
Solution. Here a = 5, r = 1/5
Tn = 0.0016 = 16/10000
16/10000 = 5. (1/5) n - 1 i.e.,
(1/5) n - 1 = (1/625) = (1/5)4
n - 2 = 4 ! n = 6. Hence 6th term is 0.0016.

Example 2. Find the 10th term of the series: 4 - 8 + 16 - 32 +...
Solution. Given a = 4 and r = -2. Thus,
Tn = ar n - 1 = T10 = ar 10 - 1 = 4 (-2)9 = - 2048.
Example3. Three numbers whose sum is 15 are in A.P. If 1, 4, and 19 are added to them respectively, the
results are in G.P. Find the numbers.
Solution. Let the three numbers in A.P. be a - d, a, a + d, so that
a - d + a + a + d = 15 or a = 5
But, numbers (a d) + 1, a + 4, (a + d) + 19 are in G.P., so that
{(a d) + 1}{(a + d) + 19} = (a + 4)2 or (5 d + 1) (5 + d +19) = (5+ 4)2
(6 d) (24 + d) = 81 or d2 + 18d 63 = 0
(d 3) (d + 21) = 0, i.e., d = 3 or d = 2.
Hence, the required numbers are
5 3, 5, 5 + 3 or 5 ( 21), 5, 5 + ( 21), i.e. 2, 5, 8 or 26, 5, 16.
Example 4. If the third term of G.P. is the cube of the first and the 6th term are 64. Find the series.
Solution. Let a be the first term and r be the common ratio. Then, the general term is given by
Tn = a rn-1.
Therefore,
T3 = a r3-1 = a r2
and
But, we are given that T3 = (T1) 3, i.e. a r2 = a3 or
T6 = a r6-1 = a r5
r=a
a r5 = 64
Also,
T6 = 64, or
or
a. a5 = 64 , a6 = 64 or a = 2
Thus, the first term is, a = 2 and common ratio r = 2. Hence, the series is 2 + 4 + 8 + .
Sum of nterms of G.P.
Let a be the first term and r the common ratio of the given G.P. of n terms and its sum to n terms be denoted
by Sn. Then,
Sn = a + ar + a r2 + a r3+.+ a rn 2 + a rn 1
.(1)
Multiplying both sides by r, we get

r. Sn = ar + a r2 + a r3+ + a r4+. + a rn 1+ a rn
Subtracting (2) from (1) we get,
24
.. (2)
(1 r) Sn = a a rn = a (1 rn)
a (1 rn)
Sn =
, for r < 1
... (3)
1r
a (rn 1)
Sn =
, for r > 1
... (4)
r1
Also, Sn = a + a + a + a + ...........to n terms = na for r = 1
Note: If l denotes the last term i.e. the nth term, then for r > 1.
a (rn 1 )
Sn =
arn a
=
r1
=
r1
a lr
Also,
lr a
arn 1.r a
.... (5)
=
r1
r1
if r < 1
.... (6)
Sn =
lr
Remarks
1. It is convenient to use formulae (3) and (6) when r < 1. and (4) and (5) when r > 1.
2. (a) If r = 1, then Sn = 0/0, which is meaningless. But in this case G.P. reduces to a, a, a and, therefore,
Sn = a + a + a + to n terms.
(b) If the number of terms in a G.P. are infinite, then
Sn =
a
1r
a
r1
when r < 1
when r > 1
Sum of an infinite G.P.

A series consisting of infinite (unlimited) number of terms is called an infinite series. In the determination of
the sum of an infinite G.P., a, ar, a r2, a r3 three cases may arise:
Sn = a (rn 1) will become Sn = , for n =
(r 1)
Case 1. If r > 1, it follows that the value of each term of series is greater than its preceding term. Then
Sn
(r n 1)
will become S n for
(r 1)
25
Case 2. If r = 1, in this case each term of the series is a, then the sum Sn for n = is given by S = a + a
+ a + to
That is, in this case also the sum will be whatever the value of a (>0) may be.
Case 3. If r < 1, in this case the value of each term will decrease with increasing value of n, because rn
0 (tends to zero) as n becomes infinitely large, i.e., r 0 when r < 1. We have
Sn = (a arn)
(1 r)
a
1r
arn
1 r
(r < 1)
When n becomes infinitely large, the quantity arn will be very small to count.
ar n
Therefore, for sufficient large value of n (i.e., = ) we can neglect the value of
1 r
Thus the sum of the infinite number of terms in G.P. is given by S
a
1 r
Solved Examples
Example1. Find the sum of the series: 1 + 5 + 25 + to 20 terms.
Solution. Given a = 1, r =5 (> 1). Hence, using the formula
Sn =
a (rn 1)
= S20 =
r1
1 (520 1)
=
1/4 (520 1)
51
Example1. Find the sum of infinite terms of the series:

a ( a + b) + a2( a2 + b2) + a3( a3 + b3) + provided a < 1 and b < 1.
Solution. We have S = a (a + b) + a2 (a2 + b2) + a3 ( a3 + b3) +
= ( a2 + ab) + ( a4 + a2b2) + ( a6+ a3b3) +
= (a2 + a4 + a6 +) + (ab + a2b2 + a3b3 +
a2
ab
, sum of inf int e G.P.' s

2
1 a 1 ab
Geometric Mean
If a, G, b are in G.P., then G is the geometric mean (G.M.) between a and b or equivalently, if
three quantities are in G.P. , then the middle one is called the G.M. between the other two.
G = ab
G
a
b
G
i.e., G2 = ab
In general, if a, G1, G2, G3, . . , Gn , b be a sequence of numbers which is a G.P., then G1, G2,
G3, . . , Gn are the n geometric means between a and b.
26
To insert n G.M.s between a and b

Suppose that G1, G2, G3, . . , Gn be the n G.M. between the two given numbers a and b.
So, a, G1, G2, G3, . . , Gn , b are in G.P.
Let r be the common ratio of the G.P. Now there are n + 2 terms in this G.P. and b is the
(n + 2)th term.
Tn + 2 = a rn + 1 = b (b/a) 1/ n + 1
G1 = a r = a. (b/a) 1/ (n + 1) G2 = a r2 = a . (b/a) 2/ (n + 1) Gn = a rn = a . (b/a) n/ (n + 1) Which are the n
G.M. between a and b.
Product of n G.M. between a and b
Here we are required to find the product of G1, G2, G3, . . , Gn.
Now,G1 G2 G3 . . Gn.
= a . (b/a) 1/ (n + 1) . a . (b/a) 2/ (n + 1) . a . (b/a) 3/ (n + 1) .. a . (b/a) n/ (n + 1)
= a n . (b/a) (1 + 2 + 3 + + n) / (n + 1) = a n. (b/a) n(n + 1) / 2(n + 1)
=(
= a n.(b/a) n / 2 = (a 1 / 2. b1/ 2) n
ab) n = nth power of the G.M. between a and b.
SOLVED EXAMPLES
Example1. Insert 3 G.M.s between
(i) 3 and 432
(ii) 1/9 and 9.
Solution. (i) Let G1, G2, G3 be the three G.M.s between 3 and 432.
So, 3 , G1, G2, G3, 432 are in G.P. Let r be the common ratio.
T5 = 3 r4 = 432
r4 = 144 = (2 3)4
Now, G1 = a r = 3. 2
r= 2 3
3 G2 = a r2 = 3.12 = 36
3=6
G3 = a r3 = 3.24 3 = 72 3 Hence the required G.M.s are 6

(ii)
3, 36, 72
3.
Here, 1/9 , G1, G2, G3, 9 are in G.P.

T5 = (1/9) r4 = 9
r4 = 81 = (3)4
r= 3
Hence the required G.M.s are ar, ar2, ar3

G1 = a r = (1/9) 3 = 1/3
G2 = a r2 = (1/9) 9 = 1 G3 = a r3 = (1/9) 27 = 3
So, the required G.M.s are 1/3, 1, 3.

Example2. The A.M. between two numbers is 20 and their G.M. is 16. Find the numbers.
Solution. Suppose the two numbers a and b.
So,
A.M. = (a + b)/ 2 = 20 ! a + b = 40
And,
G.M. =
ab = 16
.(1)
ab = 256
Now using the identity (a b) 2= (a + b) 2 4ab, we have

27
.(2)
(a b) 2= (40) 2 256
(a b) 2= 1600 1024 = 576
a b = 24.
(3)
Solving (1) and (3) , we get a = 32, b = 8 or a = 8, b = 32.

Hence, the required numbers are 8 and 32.
1.14 Applications of Arithmetic and Geometric Progressions

Example 1: For three consecutive months, a person deposits some amount of money on the first day
of each month in small savings fund. These successive amounts in the fund deposited, the total value
of which is Rs. 65, form a G.P.. If two extreme amounts be multiplied each by 3 and the mean by 5,
the product forms an A.P. Find the amounts in the first and second deposits.
Solution: Let a, ar, ar2 be successive amounts deposited. Then
a + ar + a r2 = 65
.(i)
and 3a, 5ar and 3ar2 form an A.P. Thus, by the property of A.P., we have
3a 5ar = 5ar 3ar2
3ar2 10ar + 3a = 0
3r2 10r + 3 = 0
(r 3)(3r 1) = 0, i.e., r = 3 or 1/3.
For r = 3, from (i), we have; a + 3a + 9a = 65, i.e. 13a = 65 and a = 5. Thus, the amounts are Rs. 5, Rs.
15, and Rs. 45.
Again if r = 1/3 then from (i), we have [a + (a/3) + (a/9)] = 65, i.e., a = 45. Thus, the amounts are Rs.45,
Rs. 15, and Rs. 5.
Hence, the amounts in the first and second deposits are Rs. 5, Rs. 15, and Rs. 45.
Example 2: The value of a machine depreciated each year by 20%. If the percent value of the
machine is Rs. 18,750, then find the estimated value at the end of 5 years.
Solution: (i) Depreciation for the first year = 18,750 0.20. Thus, the depreciation value of the machine at
the end of first year is : (18,750 18, 750 0.20 ) = 18,750( 1 0.20) = 15,000
(ii) Depreciation for the second year
= (Depreciation value at the end of first year) (Rate of depreciation for second year)
= 18,750 (1 0.20) (0.20)
Depreciation value at the end of the second year is:
= (Depreciated value after first year) (Depreciation for second year)
= 18,750 (1 0.20) 18,750 (1 0.20) (0.20) = 18,750 (1 0.20) 2
Similarly, depreciated value of the machine at the end of third, fourth and fifth year is:
18,750 (1 0.20) 3 18,750 (1 0.20) 4 and 18,750 (1 0.20) 5 respectively.
28
1.15. Summary
Quantitative techniques used to solve the complex problems that come in business or industrial area. For
doing this we have many techniques. We are having many techniques according to the problem. For example
to solve the problem related to profit and loss we are having profit function. For solve the problem of
demand and supply we have demand and supply function, this function decides the cost of the product
according to the demand of market, for finding revenue cost we have revenue cost function. For find out
production from a raw material we have production function. The production function helps in decision
making too. It can tell us that how many people we need to run an organization. How much capital we
needed etc. By using consumption function we can estimate our consumption. Quantitative techniques not
only help in business or economics. There are some special functions also. For example we have polynomial
function, even and odd function, rational and irrational function, one to one function, onto function, into
function, many to one, many to many, one one onto, one one into, and then we have composition of
functions. This helps us in information technology field, as well as in banks and many other areas too.
It has a very important role in todays global economy. People are doing research in this area. It is the fastest
growing technique for business sector. In every area functions helps us to solve our real problems. It also
helps in proper deployment of resources, minimizing waiting and servicing costs, choosing an optimum
strategy, optimum resources allocation, helps in decision making of management like in purchasing and
decides the price of a particular product.
1.16
Key Words
Quantitative Technique: It is a tool to help in decision making and problem solving.
Demand: Quantity demanded of a particular commodity by the buyers.
Supply: Handed particular commodity to the consumer that has been produced for them.
Cost: Fixed and variable cost at the time of producing a particular commodity.
Revenue Function: The total revenue in a particular commodity of a firm is depending on the
number of units sold.
Profit: Total Revenue Total Cost.
A.P: A sequence whose terms increase or decrease by a constant number called common difference.
G.P: A sequence whose terms increase or decrease by a constant ratio called common ratio.

1.
Describe, in brief, some of the important quantitative techniques used in modern business industrial
units.
2.
Write a short note on Operation Research describing some of the important O.R. techniques.
3.
Discuss fully the limitations of quantitative techniques.
4.
Find the domain and range of the following functions:

(i) y = 1/ (x 3)
(ii) y = 1/3x
5.
If f(x) = 2x2 + x + 1 , then simplify f(x + 1) 2 f(x) + f(x 1)
6.
Consider the quadratic equation 2x2 8x + c = 0. For what value of c the equation has
(i) real roots (ii) equal roots and (iii) imaginary roots ?
29
7.
Determine the quadratic equation whose roots are:

(i) 1, 2
(ii) a/b, b/a
(iii) 3, 2/3
8.
Find the point of equilibrium for the following demand supply equations: Q = 5 3P and Q = 4P
+12.
9.
If f(x) = log x, then proved that f(x, y) = f(x) = f(y) and f(xn) = n f(x).
10.
Graph the function y = x2 + 4x 2 with the set of values 5 x 5 as the domain.
11.
Find the quadratic function, y = a x2 + bx + c, that fits the data points (1, 4), (1, 2) and (2, 13).
Estimate the value of y when x = 3.
12.
The total cost curve for consumption of electricity is a linear function of the number of units consumed.
The cost per unit on the first 60 units is slower and on the balance over 60, is higher. If the total cost
for 100 units and 150 units consumed are Rs. 62 and Rs. 102, respectively, find the equation of the
total cost line for consumption not less than 60 units and the charge for 180 units.
13.
A manufacturer produces TV sets at a cost of Rs. 2,20,000 and 125 TV sets at a cost of 2,87,500.
Assuming the cost curve to be linear, find the equation of this line and then use it to estimate the cost
of 95 TV sets.
14.
A firm produces an item that sells for Rs. 1.75each. IT costs the firm Rs. 0.75 to produce one unit
of the item. The fixed costs are calculated to be Rs. 3000 per month. How many units would have
to be produced each year to break-even? If the total number of capacity of the plant is 72,000
units, what is the break-even in the terms of percent capacity and rupees?
15.
A television manufacturer finds that the total cost of producing television sets follows a quadratic
function given by : C(n) = 250n2 + 3250n + 10,000, where n denotes the number of television sets
produced. Each product can be marketed for Rs. 6,500. Determine the break-even point.
16.
Find the sum of the following series of the A.P.

(i) 2, 7 , 12, 17 ..
(ii) 5 2 + 1 + 4 + .to 20 terms.
(iii) 1, 2, 3,
17.
Show that the sum of (m + n)th term and (mn)th term of A.P. is equal to the twice of mth term.
18.
Find the increasing arithmetic progression, the sum of whose first three terms is 27 and sum of their
square is 275.
19.
In an A.P. which has odd number of terms. Prove that the middle term is the arithmetic mean of the
first and last term.
20.
The monthly salary of a person was Rs. 320 for each of the first three years. He next got annual
increment of Rs. 40 per month for each of the following successive 12 years. His salary remained
stationary till retirement when he found that his average monthly salary during the service period
was Rs. 698. Find the period of his service.
21.
Find (i) 6th term of the series 5, 15, 45, (ii) the nth term of the series 0.004, 0.02, 0.1,
22.
(i) If the first term of G.P. exceeds in 2nd term by 2, and the sum to infinity is 50, find the series.
(ii) The product of three numbers in G.P. is 512. IF 8 is added to the first term and 6 is added to
second, the number becomes in A.P. Find the numbers.
23.
If Sn denotes the sum of the n terms of G.P. where the first term is a and the common
r. Find the value of the sum S1 + S3 + S5 + S7 + ..S2n - 1.
30
ratio is
24.
A manufacturer records that the value of the machine costing him Rs. 8,800 depreciates at the rate
of 25% per annum. Find the estimated value at the end of 5 years.
25.
Annual increase in the population of a country is estimated to be 2%. What is the expected population
at the end of the year 1988 if it is 75 crores at the end of the year 1978?
1.18 References
Business Mathematics Theory and Applications: J.K. Sharma
A Text book of Business Mathematics: Padmalochan Hazarike
Business Mathematics: D.C. Tulsian
Quantitative Techniques: C.R. Kothari
Elements of Mathematics: Jeevansons Publications
Mathematics For Management and Economics: G. S. Monga
31
Unit - 2 : Elementary Calculus and Matrix Algebra

Unit Structure:
2.0
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
2.10
2.11
2.12
2.13
2.14
2.15
2.16
Objectives
Introduction
Functions
Limits and Continuity
Differentiation
Application of Calculus to Business
An Introduction to Matrices
Types of Matrices
Matrix Operations
Determinants
Properties of Determinants
Solution of Simultaneous Linear Equations
Application of Matrices to Business Problems
Summary
Key Words
Self Assessment Test
References
2.0
Objectives
After reading this unit, you will be able to understand

The meaning and types of functions
Limits and continuity
Application of limits to business problems (Maxima and Minima)
Meaning and importance of matrices
Types of matrices
Matrix operations
Determinants and their properties
Gauss Jordan elimination method
Application of matrices to business problems
2.1
Introduction
Calculus was developed independently by the Englishman, Sir Isaac Newton, and by the German, Gottfried
Leibniz. From Leibniz we get the dy/dx and signs.
Calculus is concerned with comparing quantities which vary in a non-linear way. It is used extensively in
science and engineering since many of the things we are studying (like velocity, acceleration, and current in
a circuit) do not behave in a simple, linear fashion.
Calculus is the study of change, in the same way that geometry is the study of shape and algebra is the study
of operations and their application to solving equations.Calculus has a strong geometric flavor. Geometric
space is nothing more than a collection of points, and geometric objects in space consist of some subcollection of points in space.
32
2.2
Functions
A function is a technical term used to symbolise relationship between variables .When two variables are so
related, that for any arbitrarily assigned value to one of them, there corresponds a definite value for the
other.The area of a circle depends upon the length of its radius and therefore we could say that area is said
to be the function of radius. The idea of function is expressed as
The relationship between two real variables say x and y which are so related that corresponding to
every value of x defined as the domain , we get a finite value b of y defined as the range then y is said
to be the function of x .
The domain of variation of x in a function is called the domain of the function Function assigns exactly one
output to each input. The argument and the value may be real numbers, but they can also be elements from
any given set. A function f with argument x is denoted f(x), which is read f of x.
The number y is called the value of the function f at x and is written as y = f(x)
An example of such a function is f(x) = 2x, the function which associates with every number x the number
twice as large. For instance, if its argument is 5 its value is 10, and this is written as:
f (5) = 10.
Types of functions:
a) Explicit and implicit functions:
The explicit function is a function expressed directly in terms of the dependent variable is said to be an
explicit function .It is denoted by: y= f(x)
Example :
Y = axn+bx where a, n and b are constant. y= 5x3-3
The Implicit function is a function which is not expressed directly in terms of of the dependent variable,
there is a mutual relationship between two variables and either variable determines the other, is said to be
implicit function .It is denoted by: R(x,y) = 0
b) Single valued and many valued functions:
A single-valued function is each element of the functions domain maps to a single, well-defined element
of its range. This contrasts with a general binary relation, which can be viewed as being a multi-valued
function.
For example: f(x) = x+3 (each element in domain has not more than one image in range set).
A multivalued function is a left-total relation, i.e. every input is associated with one or more outputs..
Strictly speaking, a well-defined function associates one, and only one, output to any particular input.
c) Even and odd functions: Function of a real variable. A function f is even if the following equation
holds for all x in the domain of f: f (x)= f(-x)
Examples of even functions are |x| , x2, x4, cos(x), and cos(x)
Again, let f(x) be a real valued function of a real variable. A function f is odd if the following equation holds
for all x in the domain of f: -f (x) = f(-x) , or f (x)+f(-x)=0
d) Parametric functions: If the value of x and y can be expressed in terms of another variable t, as x=
f(t),y=(t).
e) Continous and Discontinous functions : A function is said to be continous in a given interval (a,b)
provided if we represent the function graphically for different values, and the function does not break, it is
33
said to be continous one , otherwise discontinous.

Examples of Functions in Economics
Demand Function: If p is the price and x is the quantity of a commodity demanded by the consumers
p = f(x) or x = f(p) where q is the quantity demanded and p is the price.
Supply Function: If p is the price and x is the quantity of a commodity supplied by the firm , we write
the supply function as q = f(p) or x = f(p) where q is the quantity of a commodity supplied by a firm.
Production function: If x is the amount of output produced by the firm, when the factors or inputs land
(D), labour (l), capital (k) and organization (o)are employed . We write the production function as
q = f(D, l, k , o)where q is the amount produced by the firm.
Utility function: If U is utility of a consumer and q1 and q2, are the quantities of commodities
consumed, we write the utility function as U = f(q1, q2)
Cost Function: If x is the quantity produced by a firm at a total cost C, we write for cost function as
C= f(x)
2.3
Limits and Continuity
Limits
Definition. The limit of a function is that fixed value to which a function approaches as the variable
approaches a given value. The concept of limit is very important concept in calculus. A particular case of
a limit leads to continuity. To says that:
Means that (x) can be made as close as desired to L by making x close enough, but not equal, to p.
Continuity
A function is continuous if you can draw its graph without taking your pencil off the page. But sometimes
this will be true for some parts of a graph but not for others. Therefore, we want to start by defining what it
means for a function to be continuous at one point.
Definition: (continuity at a point)
If f(x) is defined on an open interval containing c, then f(x) is said to be continuous at c if and only if:
Note that for f to be continuous at c, the definition in effect requires three conditions:
1. that f is defined at c, so f(c) exists,
2. the limit as x approaches c exists, and
3. The limit and f(c) are equal.
If any of these do not hold then f is not continuous at c.
2.4
Differentiation
In calculus a branch of mathematics the derivative is a measure of how a function changes as its
input changes. In other words a derivative can be thought of as how much one quantity is changing in
response to changes in some other quantity; for example, the derivative of the position of a moving object
with respect to time is the objects instantaneous velocity.
34
The process of finding a derivative is called differentiation. The reverse process is called antidifferentiation. The fundamental theorem of calculus states that anti-differentiation is the same as integration.
The slope of a tangent to a curve
Applications include:
Temperature change at a particular time
Velocity of a falling object at a particular time
Current through a circuit at a particular time
Variation in stock market prices at a particular time
Population growth at a particular time
Temperature increase as density increases in a gas
Now, we will find rates of change numerically (that is, by substituting numbers in until we find an acceptable
approximation.)
We look at the general case and write our functions involving the familiar x (independent) and y (dependent)
variables.
The slope of a curve y = f(x) at the point P means the slope of the tangent at the point P. We have to find
this slope to solve many applications since it tells us the rate of change at a particular instant.
[We write y = f(x) on the curve since y is a function of x. That is, as x varies, y varies also.]
Delta Notation
In this , we write
change in y as y
change in x as x
By definition, the slope is given by:
The derivative from the first principle (Delta Method)

First principles is also known as delta method, since many texts use x (for change in x) and y (for
change in y). This makes the algebra appear more difficult, so here we use h for x instead. We still call
it delta method.
35
We wish to find an algebraic method to find the slope of y = f(x) at P, to save doing the numerical
substitutions that we saw in the last section (Slope of a Tangent to a Curve).
We can approximate this value by taking a point somewhere near to P(x, f(x)), say Q(x + h, f(x + h)).
The value g/h is an approximation to the slope of the tangent which we require.
We can also write this slope as change in y / change in x or:
If we move Q closer and closer to P, the line PQ will get closer and closer to the tangent at P and so the
slope of PQ gets closer to the slope that we want.
If we let Q go all the way to touch P (i.e. h = 0), then we would have the exact slope of the
tangent.
Now, g/h can be written:

So also, the slope PQ will be given by:
But we require the slope at P, so we let h 0 (that is let h approach 0), then in effect, Q will
approach P and g/h will approach the required slope.
Putting this to gether, we can write the slope of the tangent at P as:
This is called differentiation from first principles, (or the delta method). It gives the instantaneous rate
36
of change of y with respect to x.

This is equivalent to the following (where before we were using h for x):
You will also come across the following for delta method:
Notation for the Derivative

Important: The derivative (also called differentiation) can be written in several ways. This can cause
some confusion when we first learn about differentiation.
The following are equivalent ways of writing the first derivative of y = f(x):
or f (x) or y.
Example :
Find
from first principles if y = 2x2+ 3x.7
Solution :
f(x) = 2x2+ 3x so
We now need to find:
We have found an expression that can give us the slope of the tangent anywhere on the curve.
If x = -2, the slope is 4(-2) + 3 = -5 (red, in the graph below)
If x = 1, the slope is 4(1) + 3 = 7 (green)
If x = 4, the slope is 4(4) + 3 = 19 (black)
We can see that our answers are correct when we graph the curve (which is a parabola) and observe the
slopes of the tangents.
37
This is what makes calculus so powerful. We can find the slope anywhere on the curve (i.e. the rate of
change of the function anywhere).
Definition Via Difference Quotients
Let be a real valued function. In classical geometry, the tangent line to the graph of the function at a real
number a was the unique line through the point (a, (a)) that did not meet the graph of transversally,
meaning that the line did not pass straight through the graph. The derivative of y with respect to x at a is,
geometrically, the slope of the tangent line to the graph of at a. The slope of the tangent line is very close
to the slope of the line through (a, (a)) and a nearby point on the graph, for example (a + h, (a + h)).
These lines are called secant lines. A value of h close to zero gives a good approximation to the slope of the
tangent line, and smaller values (in absolute value) of h will, in general, give better approximations. The
slope m of the secant line is the difference between the y values of these points divided by the difference
between the x values,
that is,
This expression is Newtons difference quotient. The derivative is the value of the difference quotient as the
secant lines approach the tangent line. Formally, the derivative of the function at
a is the limit:
of the difference quotient as h approaches zero, if this limit exists. If the limit exists, then is differentiable
at a. Here 2 (a) is one of several common notations for the derivative (see below).
Equivalently, the derivative satisfies the property that:
Which has the intuitive interpretation that the tangent lines to at a gives the best linear
38
approximation:
near a (i.e., for small h). This interpretation is the easiest to generalize to other settings (see below).
Substituting 0 for h in the difference quotient causes division by zero, so the slope of the tangent line
cannot be found directly using this method. Instead, define Q(h) to be the difference quotient
as a function of h:
Q(h) is the slope of the secant line between (a, (a)) and (a + h, (a + h)). If is a continuous function,
meaning that its graph is an unbroken curve with no gaps, then Q is a continuous function away from h = 0.
If the limit exists, meaning that there is a way of choosing a value for Q(0) that makes the graph of Q a
continuous function, then the function is differentiable at a, and its derivative at a equals Q(0).
Example
The squaring function (x) = x is differentiable at x = 3, and its derivative there is 6. This result is established
by calculating the limit as h approaches zero of the difference quotient of (3):
The last expression shows that the difference quotient equals 6 + h when h 0 and is undefined when h =
0, because of the definition of the difference quotient. However, the definition of the limit says the difference
quotient does not need to be defined when h = 0. The limit is the result of letting h go to zero, meaning it is
the value that 6 + h tends to as h becomes very small:
Hence the slope of the graph of the squaring function at the point (3, 9) is 6, and so its derivative at x = 3 is
(3) = 6.
More generally, a similar computation shows that the derivative of the squaring function at x = a is (a) =
2a.
The Derivative as a Function
Let be a function that has a derivative at every point a in the domain of . Because every point a has a
derivative, there is a function that sends the point a to the derivative of at a. This function is written f2 (x)
and is called the derivative function or the derivative of . The derivative of collects all the derivatives
of at all the points in the domain of .
Sometimes has a derivative at most, but not all, points of its domain. The function whose value at a equals
f2 (a) whenever f2 (a) is defined and elsewhere is undefined is also called the derivative of . It is still a
function, but its domain is strictly smaller than the domain of .
The derivative is an operator whose domain is the set of all functions that have derivatives at every point of
their domain and whose range is a set of functions. If we denote this operator by D, then D () is the function
f2 (x). Since D () is a function, it can be evaluated at a point a. By the definition of the derivative function,
D () (a) = f2 (a).
For comparison, consider the doubling function (x) =2x; is a real-valued function of a real number,
39
meaning that it takes numbers as inputs and has numbers as outputs:
The operator D, however, is not defined on individual numbers. It is only defined on functions:
Because the output of D is a function, the output of D can be evaluated at a point. For instance,
when D is applied to the squaring function
D outputs the doubling function,
which we named (x). This output function can then be evaluated to get (1) = 2, (2) = 4, and so on.
Inflection Point
A point where the second derivative of a function changes sign is called an inflection point. At an inflection
point, the second derivative may be zero, as in the case of the inflection point x=0 of the function y=x3, or
it may fail to exist, as in the case of the inflection point x=0 of the function y=x1/3. At an inflection point, a
function switches from being a convex function to being a concave function or vice versa.
Computing the Derivative
The derivative of a function can, in principle, be computed from the definition by considering the difference
quotient, and computing its limit. In practice, once the derivatives of a few simple functions are known, the
derivatives of other functions are more easily computed using rules for obtaining derivatives of more
complicated functions from simpler ones.
Derivatives of Elementary Functions
Most derivative computations eventually require taking the derivative of some common functions. The
following incomplete list gives some of the most frequently used functions of a single real variable and their
derivatives.
Derivatives of powers: if
where r is any real number, then
wherever this function is defined. For example, if f(x) = x1 / 4, then

Exponential and logarithmic functions:
40
Rules for Finding the Derivative
In many cases, complicated limit calculations by direct application of Newtons difference quotient can be
avoided using differentiation rules. Some of the most basic rules are the following.
Sum rule:
Here, u and v are functions of x. The derivative of the sum is equal to the derivative of the first plus
derivative of the second.
Product rule:
If u and v are two functions of x, then the derivative of the product uv is given by...
In words, this can be remembered as:

The derivative of a product of two functions is the first times the derivative of the second, plus the second
times the derivative of the first.
Example: If u and v are two functions of x, then the derivative of the product uv is given by...
If we have a product like
y = (2x2 + 6x) (2x3 + 5x2)
we can find the derivative without multiplying out the expression on the right.
Solution: We use the substitutions u = 2x2 + 6x and v = 2x3 + 5x2.
We can then use the product rule:
We first find:
and
Then we can write:
Quotient rule:
(A quotient is just a fraction.)

If u and v are two functions of x, then the derivative of the quotient u/v is given by...
41
In words, this can be remembered as:

The derivative of a quotient equals bottom times derivative of top minus top times derivative of the bottom,
divided by bottom squared.
Example: We wish to find the derivative of the expression:
Solution: We recognize that it is in the form: y = u/v.
We can use the substitutions:
u = 2x3 and v = 4 x
Using the quotient rule, we first need to find:
And
then
Chain rule: It is a formula for computing the derivative of the composition of two or more
functions.
f '(x) =
Example: Find (3x)
2.5
Application of Calculus to Business
Lets take a look at some applications of derivatives in the business world. For the most part these are
really applications that weve already looked at, but they are now going to be approached with an eye
towards the business world.
Lets start things out with a couple of optimization problems. Weve already looked at more than a few of
these in previous sections so there really isnt anything all that new here except for the fact that they are
coming out of the business world.
Example 1. An apartment complex has 250 apartments to rent. If they rent x apartments then their
monthly profit, in dollars, is given by,
How many apartments should they rent in order to maximize their profit?
42
Solution. All that were really being asked to do here is to maximize the profit subject to the constraint that
x must be in the range.
First, well need the derivative and the critical point(s) that fall in the range. 0 x 250
The derivative p(x) of the the profit equation p(x) is as follows:
P(x) = -16x+3200
= 3200 -16x = 0
= 3200/16
= 200
Since the profit function is continuous and we have an interval with finite bounds we can find the maximum
value by simply plugging in the only critical point that we have (which nicely enough in the range of acceptable
answers) and the end points of the range.
Substituting values in the original function i.e..
P (0) = 80,000
= 2,40,000
P (250) = 2,20,000
So, it looks like they will generate the most profit if they only rent out 200 of the apartments instead of all
250 of them.
Example 2. A production facility is capable of producing 60,000 pens in a day and the total daily cost of
producing pens x in a day is given by,
How many pens per day should they produce in order to minimize production costs?
Solution
Here we need to minimize the cost subject to the constraint that x must be in the range .
Note that in this case the cost function is not continuous at the left endpoint and so we wont be able to just
plug critical points and endpoints into the cost function to find the minimum value.
Lets get the first couple of derivatives of the cost function.
The critical points of the cost function are,
Now, clearly the negative value doesnt make any sense in this setting and so we have a single critical point
in the range of possible solutions: 50,000.
Now, as long as the second derivative is positive and so, in the range of possible solutions the function is
always concave up and so producing 50,000 pens will yield the absolute minimum production cost.
Example 3. The production costs per day for some widget is given by,
43
What is the marginal cost when , and ?

Solution
So, we need the derivative and then well need to compute some values of the derivative.
So, in order to produce the 201st widget it will cost approximately $10. To produce the 301st widget will
cost around $38. Finally, to product the 401st widget it will cost approximately $78.
2.6
An Introduction to Matrices
The knowledge of matrices is required in many areas of business and other spheres of life.In this section we
shall study the fundamentals of matrix algebra.
Suppose we wish to express that Shyam has 10 books. We may express it as (10) with the understanding
that the number inside ( ) denotes the number of books that Shyam has. Next suppose we want to express
that Shyam has 10 books and 5 pens . We may express it as (10 5) with the understanding that the first entry
inside ( ) denotes the number of books while the second entry, the number of pens, possessed by Shyam .
Let us now suppose that we have three children Shyam , Krishna and Mohan . Suppose Shyam has 10
books, 5 pens, Krishna has 4 books, 6 pens and Mohan has 5 books, 8 pens. This information can be
represented in the tabular form as follows:
Books
Pens
Shyam
10
Krishna
Mohan
We can also briefly write these as follows:

First Column
Second column
First Row
10
Second Row
Third row
The Information behind can be expresse3d in two rows and three columns as given below:
Shyam
Krishna
Mohan
Books
10
Pens
Definition:
A matrix is an array of numbers. A matrix with m rows and n columns is order m x n and is shown as
follows..
44
2.7
Types of Matrices
In this section we are going to discuss about types of matrices concept. In mathematics, a matrix is a
rectangular array of numbers, such as
An item in a matrix is called an entry or an element. The example has entries 1, 9, 13, 20, 55, and 4. Entries
are often denoted by a variable with two subscripts, as shown on the right. Matrices of the same size can be
added and subtracted entry wise and matrices of compatible sizes can be multiplied.
A row matrix has one row of numbers as shown below:-
A column matrix has one column of numbers as shown below:-
A square matrix is one with an equal number of rows and columns i.e m = n
For example,
5
A diagonal matrix is a square matrix with all numbers zero apart from diagonal numbers
as shown below:-
A unit matrix (also called as Identity Matrix) is a square matrix with all diagonal
numbers = 1-. The other elements being 0..
A null matrix/Zero Matrix has all elements = 0
For example,
0
45
2.8
Matrix Operations
In this section, we shall define equality of matrices, multiplication of a matrix by a scalar , addition, subtraction
and multiplication of matrices.
Equality of matrices
Definition: Two matrices A= (aij) and B= (b ij ) are said to be equal if
i)
ii)
they are of same order and

each elem ent of A is equal to the corresponding elem ent of B i.e., aij = b ij for
all i and j
For example, if
-1
Then a =5, b =-1, c =1, d =0, e = 3, f = 4

For Example
If
x-y
2x -y
-1
Find x, y, z, w.
Solution. By definition of equality of matrices , we have
x-y = -1
2x-y = 0
z=4
w=5
Solving (1) and (2), we get
(1)
(2)
(3)
(4)
X = 1, y = 2
Hence x = 1, y = 2, z = 4, w = 5
Multiplication of a matrix by a scalar
Definition: If A = (aij) is a matrix and k is a scalar , then kA is another matrix whose ( i, j)th element is k
aij for all possible values of i and j.
Thus, k A = (k aij)
Example:
If A =
-2
-1
, find 3A.
46
Solution. We have
3A = 3
-2
-1
-6
12
15
-3
18
Addition of Matrices
Definition: The expression zij = aij + bij means to element in row i, column j of matrix A add element in
row i, column j of matrix B. If we do this with each element of A and B we end with matrix Z. An example
is given in
Figure
The above figure shows addition operation ,in the same manner the subtraction is performed in analogous
manner. The expression zij = aij - bij, means to element in row i, column j of matrix A deduct element in row
i, column j of matrix B. If we do this with each element of A and B we end with matrix Z.
Subtraction of Matrices
In the above ex ample if we want to find the subtraction between the two m atrices we
will perform the following opera tion:
A=
B=
47
A-B =
Multiplication of Matrices
Definition:Consider two matrices A and B with the following characteristics: the number of columns in A
equals the number of rows in B. These are conformable with respect to one another, and they can be
multiplied together to form a new matrix Z.
The expression zij = ai1* b1j + ai2* b2j + ai3* b3j + ... aim* bnj, means add the products obtained by
multiplying elements in each i row of matrix A by elements in each j column of matrix B. Figure 2.6 shows
what we mean by this statement.
Figure
The order in which we multiply terms does matter. The reason for this is that we need to multiply row
elements by column elements and one by one. Therefore A*B and B*A can produce different results.
We say can produce because there exist special cases in which the operation is commutative (order
does not matter). .
2.9
Determinants
The determinant of a square matrix A = [ai j] is a number denoted by |A| or det(A), through which
important properties such as singularity can be briefly characterized. This number is defined as the following
function of the matrix elements:
|A| = det (A) = a1 j1 a2 j2 . . . an jn,
where the column indices j1, j2, . . . jn are taken from the set {1, 2, . . . n}, with no repetitions allowed. The
plus (minus) sign is taken if the permutation ( j1 j2 . . . jn) is even (odd).
To indicate that we are referring to determinant A and not to matrix A we surround the symbol A by pipes
48
(|).
Determinants are like matrices, but done up in absolute-value bars instead of square brackets. There is a lot
that you can do with (and learn from) determinants, but youll need to wait for an advanced course to learn
about them.
Determinant Order.
Determinant of a 2 by 2 Matrix
The determinant of the 2 by 2 matrix A
A=
is a scalar given by
Example : Find the determinant of the 2 by 2 matrix A given by
Solution:
det(A) = -2*(1/2)-5*3
=-16
Determinant of a 3 by 3 Matrix
T h e d eter m ina nt o f th e 3 b y 3 m atrix A
is a scalar g iv en b y
E x am p le : Find the d eterm in an t o f the 3 b y 3 m atr ix A g iv en b y
S o lu tio n
d et ( A ) = 2 ( 5 *5 -4 * 1 ) - (- 1) * (- 2 *5 - 4* 0 ) + 3* ( -2 * 1 - 5 * 0) = 4 2 - 1 0 -6 = 26
49
Expansion using Minors and Cofactors

The definition of determinant that we have so far is only for a 22 matrix. There is a shortcut for a 33
matrix, but I firmly believe you should learn the way that will work for all sizes, not just a special case for
a 33 matrix.
The method is called expansion using minors and cofactors. Before we can use them, we need to define
them.
Minors
A minor for any element is the determinant that results when the row and column that element are in are
deleted.
The notation Mij is used to stand for the minor of the element in row i and column j. So M21 would mean
the minor for the element in row 2, column 1.
Consider the 33 determinant shown below. Ive included headers so that you can keep the rows and
columns straight, but you wouldnt normally include those. Were going to find some of the minors
C , C2 , C3.
Finding the Minor for R2C1
The minor is the determinant that remains when you delete the row and column of the element youre
trying to find the minor for. That means we should delete row 2 and column 1 and then find the
determinant.
C1 C3
R1 3 2
R3 5
= 3(2)-5(2) = 6-10 = -4
As you can see, the minor for row 2 and column 1 is M21 = -4.
Lets try another one.
Finding the Minor for R3C2
This time, we would delete row 3 and column 2.
C1 C3
R1 1 2
R2 4 3 = 1(3) 4(2) = 3 8 = -5
So the minor for row 3, column 2 is M32 = -5.
Co-Factor of Determinant element
Associated with each element is a sign (+ or -) this is established by counting the number of row and column
steps to get from the first element and if the steps are 0 or even the sign is (+ ) and if the steps are odd the
sign is ( - ). e.g.
The co-factor of and element is the minor of the element x the relevant sign. e.g.
50
Expansion of Determinant
The expansion of a determinant is completed by adding the product of all elements with their co-factors for
any row or column. eg
An example of the expansion of a determinant is provided below..
2.10
Properties of Determinants
In this section, we will study properties determinants have and we will see how these properties can help in
computing the determinant of a matrix. We will also see how these properties can give us information about
matrices.
t
1.|A |= |A|
The determinant of matrix A and its transpose At are equal.
A = A* = -2
2. |A|= 0 If:
(i) It has two equal lines
(ii) All elements of a line are zero.
(iii) The elements of a line are a linear combination of the others.
r3 = r1 + r2
51
3.A triangular determinant is the product of the diagonal elements.
4. If a determinant switches two parallel lines its determinant changes sign.
5. If the elements of a line are added to the elements of another parallel line previously multiplied by a real
number, the value of the determinant is unchanged.
6. If a determinant is multiplied by a real number, any line can be multiplied by the above mentioned number,
but only one.
7. If all the elements of a line or column are formed by two addends, the above mentioned determinant
decomposes in the sum of two determinants.
8. |AB| =|A||B|
The determinant of a product equals the product of the determinants.
Examples
1. Apply the properties of determinants and calculate:
52
2. Apply the properties of determinants and calculate:
2.11 Solution of Simultaneous Linear Equations

In mathematics, simultaneous equations are a set of equations containing multiple variables. This set is
often referred to as a system of equations. A solution to a system of equations is a particular specification
of the values of all variables that simultaneously satisfies all of the equations. To find a solution, the solver
needs to use the provided equations to find the exact value of each variable. Generally, the solver uses either
a graphical method, the matrix method, the substitution method, or the elimination method. This is a set of
linear equations, also known as a linear system of equations: 2x + y = 8 and x + y = 6
Solving this involves subtracting x + y = 6 from 2x + y = 8 (using the elimination method) to remove the yvariable, then simplifying the resulting equation to find the value of x, then substituting the x-value into either
equation to find y.
The solution of this system is: x = 2 and y = 4
which can also be written as an ordered pair (2, 4), representing on a graph the coordinates of the point of
intersection of the two lines represented by the equations.
53
An example of solving 3 linear equations can be expressed as

a1x + b1y+ c1y + d1 = 0
a2x + b2y+ c2y + d2 = 0
a3x + b3y+ c3y + d3 = 0
Using determinants this is solved by the following relationship
Example:
Solve the equation.
5x - 6y + 4z = 15
7x + 4y - 3z = 19
2x + y + 6z = 46
This equation is rewritten as
5x-6y+4z-15=0
7x+4y-3z-19=0
2x + y + 6z - 46 = 0
Expressing this in determinant form
On evaluating the denominators.

( x /-1257 ) = ( y /-1676 ) = ( z / -2514 ) = ( -1 /419 )
Dividing each denominator by 419 results in
( x /3 ) = ( y /-4 ) = ( z / -6 ) = -1
These results in
x = 3: y =4: z =6
The expansion of the determinant is simplified by using the row or column with zero elements if any as the
relevant element-co-factor is zero..
Apart from the above solution there are three methods to show that how simultaneous linear equations can
be solved
Three methods:
1) Gauss Jordan Elimination /Do little Method
In linear algebra, GaussJordan elimination is an algorithm for getting matrices in reduced row echelon
form using elementary row operations. It is a variation of Gaussian elimination. Gaussian elimination places
54
zeros below each pivot in the matrix, starting with the top row and working downwards. Matrices containing
zeros below each pivot are said to be in row echelon form. GaussJordan elimination goes a step further by
placing zeros above and below each pivot; such matrices are said to be in reduced row echelon form. Every
matrix has a reduced row echelon form, and GaussJordan elimination is guaranteed to find it.
It is named after Carl Friedrich Gauss and Wilhelm Jordan because it is a variation of Gaussian elimination
as Jordan described in 1887. However, the method also appears in an article by Clasen published in the
same year. Jordan and Clasen probably discovered GaussJordan elimination independently.
Gauss-Jordan reduction
Step 1: Form the augmented matrix corresponding to the system of linear equations.
Step 2: Transform the augmented matrix to the matrix in reduced row echelon form via elementary row
operations.
Step 3: Solve the linear system corresponding to the matrix in reduced row echelon form. The solution(s)
are also for the system of linear equations in step 1.
Example:
Solve the following linear system:
Solution:
The Gauss-Jordan reduction is as follows:
Step 1:
The augmented matrix is
1
2
1
5
1
3
2
1
1
2
5
9
3
7
3
3 .
11
Step 2:
After elementary row operations, the matrix in reduced row echelon form is
1
0
5
2 .
3
Step 3:
The linear system corresponding to the matrix in reduced row echelon form is
x1
2 x4 5
x2
The solutions are
3x4 2
x3 2x4 3
55
x1 5 2t , x 2 2 3t, x 3 3 2t , x 4 t , t R
x1 5 2 t 5 2
x

2 3t 2 3
2
t
x 3 3 2t 3 2

t
x4
0 1
Number of solutions of a system of linear equations:

For any system of linear equations, precisely one of the following is true.
I.
The system has exactly one solution.
II.
The system has an infinite number of solutions.
III. The system has no solution.

Note:
The linear system with at least one solution is called consistent and the linear system with no
solution is called inconsistent.
2) Cramers Rule
Cramers rule is a method of solving a system of linear equations through the use of determinants.
Cramers rule is given by the equation
xi =
Ai
A
where xi is the i th endogenous variable in a system of equations, A is the determinant of the original
A matrix as discussed in the previous section, and Ai is the determinant a special matrix formed as
part of Cramers rule.
To use Cramers rule, two (or more) linear equations are arranged in the matrix form A x = d. For a two
equation model:
A
x = d
a11 a12 x1
d1
a21 a22 x2 = d2
A is the matrix corresponding to the number of equations in a system (here, two equations), and the number
of endogenous variables in the system (here 2 variables). Remember that the matrix must be square, so the
number of equations must equal the same number of endogenous variables. Position x has one column and
corresponds to the number of endogenous variables in the system. Finally, position d contains the exogenous
terms of each linear equation.
Note: The determinant for a matrix must not equal 0 (A 0). If A = 0 then there is no solution, or
there are infinite solutions (from dividing by zero). Therefore, A 0. When A 0, then a unique
solution exists.
Applying Cramers Rule in a 2x2 example
Using Cramers rule to solve for the unknowns in the following linear equations:
2x1 + 6x2 = 22
-x1 + 5x2 = 53
56
Then,
2 6
x1
-1 5
x2
d
22
=
53
2 6
The primary determinant A =
= 2 (5) - (-1) 6 = 16
-1 5
We need to construct
xi = Ai, for i=1 and for i=2.

A
The first special determinant A1 is found by replacing the first column of the primary matrix with the constant
d column. The new special matrix A1 now appears as:
22 6
A1 =
53 5
and solved as a regular matrix determinant,
A1 = 22 (5) - 53 (6) = -208
Likewise, the same procedure is done to find the second special determinant A2,
2 22
A2 =
-1 53
A2 = 2 (53) - (-1) (22) = 128

We have now determined:
A = 16
A1 = -208
A2 = 128
Using:
xi =
Ai
A
We get,
A1
x1 = A =
x2 =
A2
A
-208
16
128
= 16
= -13
= 8
57
Applying Cramers Rule in a 3x3 example

Using Cramers Rule to solve for the unknowns in three linear equations:
5x1 - 2x-2 + 3x3 = 16
2x1 + 3x2 - 5x-3 = 2
4x1 - 5x-2 + 6x3 = 7
Then,
5 -2 3
2 3 -5
4 -5 6
x1
16
x2 = 2
x3
7
The primary determinant A=
5 -2 3
2 3 -5 = 5(18 - 25) + 2(12 + 20) + 3(-10 - 12) = - 37
4 -5 6
The three special determinants are:
A1 =
16 -2 3
2 3 -5
7 -5 6
A2 =
5 16 3
2 2 -5 = 5(12 + 35) - 16(12 + 20) + 3(14 - 8) = -259
4 7 6
A3 =
5 -2 16
2 3 2 = 5(21 + 10) + 2(14 - 8) + 16(-10 - 12) = -185
4 -5 7
= 16(18 - 25) + 2(12 + 35) + 3(-10 - 21) = -111
Applying Cramers Rule:

x1 =
A1 -111
A
= -37 = 3
x2 =
A2
-259
A
= -37 = 7
x3 =
A3
A
-185
= -37 = 5
3) Matrix Inverse Method:

The transpose of a matrix is often referenced, but what does is mean? It sure has an algebraic interpretation
but we do not know if that could be expressed in just a few words. Anyway,we define it by taking examples
to find out what the pattern is. Below is a 2x2 matrix like it is used in complex multiplication. The transpose
of a square matrix can be considered a mirrored version of it: mirrored over the main diagonal. That is the
diagonal with the s on it.
58
Example,
-3
Find transpose of the matrix A =

And verify that (A) = A
Solution. From the definition of the transpose of a matrix, A is obtained by interchanging the rows
and columns of the matrix A. Thus
A =
-3
Further , taking the transpose of the matrix A , we have
(A) =
-3
=A
Hence (A) = A
The following transpose properties are observed in matrices

(ABC)T = CTBTAT
(ABCT)T = (CT)TBTAT = CBTAT
The following inverse properties are observed in matrices
(ABC)-1 =C-1B-1A-1
(ABC-1)-1 = (C-1)-1B-1A-1 = CB-1A-1
A-1A = AA-1 = I = 1
Since matrix division is not defined, it is impossible to divide a matrix expression by a given matrix.
Finding out Adjoint of a Matrix :
Adjoint or Adjugate Matrix of a square matrix is the transpose of the matrix formed by the cofactors of
elements of determinant |A|.
To calculate adjoint of matrix we have to follow the procedure
a) Calculate Minor for each element of the matrix.
b) Form Cofactor matrix from the minors calculated.
c)Finding cofactors from Minors
d) Form Adjoint from cofactor matrix.
For example we will use a matrix
A11
a12
a13
A21
a22
a23
A31
a32
a33
A=
59
Step 1: Calculate Minor for each element.

To calculate the minor for an element we have to use the elements that do not fall in the same row
and column of the minor element.
Minor of A11 = M11 =
A11
A12
A13
A21
A22
A23
A31
A32
A33
A11
A12
A13
A21
A22
A23
A31
A32
A33
A11
A12
A13
A22
A23
A32
A33
A21
A23
A31
A33
A21
A22
=A22 xA33 -A32 xA23
=A21 xA33 -A31 xA23
=
A21
A22
A23
A31
A32
A33
A11
A12
A13
A21
A22
A23
A31
A32
A33
=A21 xA32-a31 xA22

A31
A32
A12
A13
A32
A33
Similarly
M22 = A11 xA33 A31 xa13
M23 = A11 xA32 A31 xA12
M31= A12 xA23 A22 xA13
M32=A11 xA23 A21 xA13
Step 2: Form a matrix with the minors calcu lated.
Matrix of Minors =
M11
M12
M31
M21
M22
M23
M31
M32
M33
Step 3: Finding the cofactor from Minors:

Cofactor:A signed minor is called cofactor.
The cofactor of the element in the ith row, jth column is denoted by Cij
Cij = (-1)i+j Mij
60
= A12 xA33-A32 xA13
M atrix of Cofactors = (-1)1+1M 11
(-1) 1+ 2 M 12
(-!) 1+ 3 M13
(-1) 2+1 M 21
(-1) 2+ 2M 22
(-1) 2+3 M23
(-1) 3+1 M 31
(-1)3+2M 32
(-1)3+ 3M33
So,
C 11
C 12
C 31
C 21
C 22
C 23
C 31
C 32
C 33
M 11
-M 12
M3 1
-M 21
M 22
-M2 3
M 31
-M 32
M3 3
Step 4: Calculate adjoint of matrix:

To calculate adjoint of matrix, just put the elements in rows to columns in the cofactor matrix. i.e convert
the elements in first row to first column, second row to second column, third row to third column.
C11
C12
C31
C21
C22
C23
C31
C32
C33
Adjoint of Matrix =
So we can say that,

A-1 =
(adjoint of A) or A-1 =
(cofactor matrix of A) T
Example: The following steps result in A-1 for
The cofactor matrix for A is

we get
, so the adjoint is
. Since det A = 22,
2.12 Application of Matrices to Business Problems

We can measure the expansion of the world by matrices cause in magnetic fields vectors can be stretched
up to a certain limit which are the eigen values. It enables students to interact with numbers. Matrices are
used basically to solve linear equations....... Equations are generally solved to get value of unknown
variables.......Variable values are calculated (or assumed) to know all working. varied applications. A few
are: 1. Which group of customers are more valuable (profitable) to the business 2. How vary the business
results are from the average 4. Predict the business future.
61
The thing about business applications is that they can usually be run on equipment over five years old.
They simply do not demand much in the way of features or power.
Example
1) Ram Shyam and Mohan purchased biscuits of different brands P, Q and R . Ram purchased 10 packets
of P, 7 packets of Q and 3 packets of R. Shyam purchased 4 packets of P, 8 packets of Q and 10 packets
of R.Mohan purchased 4 packets of P, 7 packets of Q and 8 packets of R. If brand P costs Rs .4, Q costs
Rs. 5 and R costs Rs. 6 each , then using matrix operations find the amount of money spent by these persons
individually.
Solution.
Let Q be the matrix denoting the quantity of each brand of biscuit bought by P, Q and R and let C be the
matrix showing the cost of each brand of biscuit
P
Ram
10
Shaym
10
Mohan
=
C=
Since number of columns of first matrix should be equal to the number of rows of the second
matrix for multiplication to be possible, the above matrices shall be multiplied in the following
order
Q* C =
10
93
10
116
99
Amount spent by Ram, Shyam and Mohan is Rs 99, Rs 116 and Rs 99 respectively
2.13 Summary
Calculus has two main parts: differential calculus and integral calculus. Differential calculus studies the derivative
and integral calculus studies (surprise!) the integral. The derivative and integral are linked in that way that
they are both defined via the concept of the limit: they are inverse operations of each other (a fact sometimes
known as the fundamental theorem of calculus): and they are both fundamental to much of modern science
as we know it.
62
Derivatives can be used to help you graph functions. First, they give you the slope of the graph at a point,
which is useful. Second, the points where the slope of the graph is horizontal (f(x) = 0) are particularly
important, because these are the only points at which a relative minimum or maximum can occur (in a
differentiable function).
Methods of computing determinants:
(i)
Expand along a row or down a column
(ii)
Simplify using elementary row (or column) operations to get to upper triangular form.
(iii)
A mixture of (i) and (ii), i.e. use some row (or column) operations to simplify before expanding
whats left.
Methods of computing A1:

(i)
Standard formula in 2 2 case;
(ii)
Adjugate matrix method: A1 = (detA)1adjA.

Too hard for large n, but easy for n = 2 (gives formula referred to in (i)), and still possible for n = 3.
(iii)
Elementary row operations on augmented matrix (Gaussian elimination). Best general method for
large n.
Methods of solving AX = B:
(i)
Inverse matrix method: X = A1B.

Depends on knowing A-1 good way if A-1 is already known or is easy tofind, and especially
good if there are several Bs with only one A.
(ii)
Elementary row operations (or Gaussian elimination). Best general method, especially for large n.
(iii)
Cramers rule: xi = detBi/ detA, where Bi is formed from A by substituting B for ith column.
2.14 Key Words
Function : A function f is a relation from a non empty set A into a non empty set B such that
domain of f is A and no two ordered pairs in f have the same first component. Equivalently, a
function f from a set A into set B is a relation from A into B if for each aB such that
First derivative: If the first derivative is positive, the function is increasing; if the first derivative is
negative, the function is decreasing.If f (a) = 0 then a is called a critical point of f.
Second derivative: If f (a) = 0 then a is called an inflection point of f.
Explicit function is a function in which the dependent variable has been given explicitly
Differentiation: is defined as a method to compute the rate at which a dependent output y

changes with respect to the change in the independent input x.
Crammers Rule: The linear system AX = B has a unique solution if and only if A is invertible. In
this case, the solution is given by the so-called Cramers formulas
Co-Factor of Determinant:- Associated with each element is a sign (+ or -) this is established

by counting the number of row and column steps to get from the first element and if the steps are
0 or even the sign is (+ ) and if the steps are odd the sign is ( - ).
63

1
There are three dealers in a market who sales Radios, T.V and CD players . A sells weekly 14
radios, 2 TV and CD players. B sells weekly 10 radios, 3 TV and 10 CD players. C sells weekly
9 radios, 5 TV and 9 CD players. If the profit per radio is Rs. 50, on TV is Rs. 20 and on CD
players is Rs. 30. Calculate their individual profits.
A company has two plants P1 and P2, P1 is capable of producing 5 units of A, 10 units of B and 3
units of C per hour of operation. P2 is capable of producing 5 units of A, 6 units of B and 6 units of
C per hour of operation. Using matrix algebra, determine the total number of units of A, B and C
produced if P1 is operated for 10 hrs and P2 is operated for 5 hours.
A man invested Rs. 30000 into three different investments. The rates of interest is 2%, 3% and 4%
per annum resp. The total annual income is Rs. 1000. If the income from the first and second
investments is Rs. 50 more than the income from third, find the amount of each investment, by using
matrix algebra.
Vitamin A and B are found in two different foods F1 and F2. One unit of food F1 contains 3 units
of vit A and 8 units of vitamin B. One unit of food F2 contains 4 units of vitamin A and 6 units of
vitamin B. The daily requirements of vitamin A and B is 24 and 48 units resp. Calculate the optimum
mixture of food F1 and F2 to be taken to meet the daily requirements of vitamin A and B.
5 Show that
13
16
19
14
17
20
15
18
21
=0
Solve the following system of linear equations using the method of matrix reduction
x+y+z=6
x + 2y + 3z = 14
-x + y - z = -2
For square matrices A and B , expand (A + B) (A B) and (A B) (A + B). When will these be
equal?
2.16 References
Mathematics for management and economics by G.S. Mongia,Vikas Publishing House.

A text book of Business Mathematics by Kavita Gupta,Sun India Publications.
Business Mathematics by J.K Thukral,Mayur Paperbacks.
A Text book of Business Mathematics by J.K Sharma,Ane Books Pvt. Ltd.
Business Mathematics with Applications by S.R Arora and Dinesh Khattar,S.Chand Publishing
House.
Mathematics for Management (An introduction ) by M Raghavachari,Tata McGraw-Hill
Publishing Company Limited.
wiki.answers.com
www.scribd.com
ndu2009algebra.blogspot.com
www.enotes.com
Business Mathematics by D.R.Agarwal.
64
Unit - 3 : Correlation and Simple Regression

Unit Structure:
3.0
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
3.10
Objectives
Introduction
Correlation
Correlation Coefficient
Simple Linear Regression
Estimating Linear Regression
Difference between Correlation and Regression
Summary
Key Words
References
3.0
Objectives
After studying this unit, you will be able to
Define Correlation and Regression.
Differentiate between Correlation and Regression.
Find the equation of the lines of Regression.
Find out how much a variable changes corresponding to the average amount of change in the other
variables.
3.1
Introduction
From statistics point of view Correlation and Regression are the tools to measure the amount of similarity
and variations of two or more variables. These are used for studying and measuring the extent of relationship
between two or more variables. Generally, there are situations where there is a certain degree of association
but not of a fixed type between variables which may occurs in pair or groups. For example, consider the
heights and weights of 100 students in a class. We want to measure the heights and weights of all the
students. We observe that if the height and weight of a student is represented by x and y respectively, then
there corresponds a pair of values (xi,yi) of the variables x and y for each measurement. Such a distribution
in which each individual of the set is made up of two values is called bivariate distribution.
If the changes in the value of one variable appear to be related to the changes in the value of another
variable, then the two variables are said to be correlated. For example, pressure and volume of a gas are
correlated because with the increase in pressure there is decrease in volume and the vice versa. Other
similarly placed situations are:
The amount of yield of grain per acre and the amount of fertilizer per acre.
The sales revenue and advertising expenditure of a firm in number of years.
The heights of the husband and the wives at the time of marriage.
65
In a bivariate distribution, we intend to find a relationship (if it exist) between two variables under study.
Francis Galton used the term Regression in the latter part of nineteenth century. Regression is used to find
out some sort of functional relationship between two or more variables. The average relationship between
the correlated variables is estimated by Regression. The main objective of correlation is to study the relation
(if it exists) between two or more variables which are so related that a change in one variable is accompanied
by changes in other variable. Regression, on the other hand, aims at estimation of unknown values of one
variable from the known values of other variable.
3.2
Correlation
We define Correlation as a measure of relation between two variables. In other words, to describe and
understand the association between two continuous variables (interval or ratio data), we compute Correlation.
Thus two variables are said to be correlated if an increase (or decrease) in one variable is accompanied by
an increase or decrease (decrease or increase) in the other variable.
Types of Correlation
Correlation is classified in several different ways:
1. Positive or Negative correlation
2. Linear and Non-linear correlation
3. Simple, Multiple and Partial correlation
4. Spurious correlation
Positive Correlation
Two variables are said to be positively correlated, if an increase in the value of one is accompanied by an
increase in the value of other or a decrease in the value of one is accompanied by a decrease in the value of
other i.e. the value of two variables deviate in the same direction. For example, paired variables like supply
and demand of commodities, household income and expenditure, price and supply of commodities is positively
correlated.
Examples:
X:
10
12
14
16
18
Y:
15
20
25
30
35
X:
80
75
70
60
50
Y:
Negative Correlation
Two variables are said to be negatively correlated, if an increase in the value of one is accompanied by a
decrease in the value of other or a decrease in the value of one is accompanied by an increase in the value
of other. For example, paired variables like pressure and volume of a gas, current and resistance (voltage
being constant), demand and price of commodities are negatively correlated.
66
Examples:
X:
10
12
14
16
18
Y:
15
12
10
X:
80
75
70
60
50
Y:
60
70
75
80
90
Linear Correlation
The correlation between two variables is said to be linear, if there exists a relationship of the form:
y = a + bx
where, a and b are real numbers.
In a linear correlation the amount of change in one variable tends to bear constant ratio to the amount of
change in the other variable. The graph of linear correlation is a straight line.
Example:
X:
10
12
14
16
18
Y:
15
20
25
30
35
Non Linear Correlation

The correlation between two variables is said to be non - linear, if there exists a relationship of the form of
polynomial of order more than one, for example:
y = a + bx + cx2
where, a , b and c are real numbers.
In the non - linear correlation, a change of one unit in one variable does not correspond to same amount
of change in the other variable. The graph of non - linear correlation is not a straight line.
Example:
X:
10
12
14
16
18
Y:
15
20
22
30
40
Simple Correlation
In simple correlation the number of variables studied/ included is two. All the above mentioned correlations
are simple; including positive, negative, linear and non-linear.
Multiple Correlation
When three or more variables are studied simultaneously it is a case of multiple correlation. For example,
when we study the relationship between the yield of rice per acre, amount of rainfall and the amount of
fertilizers used, it is a problem of multiple correlation.
67
Partial Correlation
In partial correlation we recognize more than two variables, but consider only two variables to be influencing
each other the effect of other influencing variables being kept constant. For example, in the rice problem
taken above if we limit our correlation analysis of yield and rainfall to periods when a certain average daily
temperature existed it becomes a problem relating to partial correlation only.
Spurious correlation
Given data regarding any two variables, it is possible that on calculation of r, one may say that statistically
the two variables are correlated, but if there is no justifiable, logical explanation for correlation to exists, then
such statistical correlation is termed as spurious correlation. In a spurious correlation two variables have
no direct causal connection, yet it may be wrongly inferred that they do, due to either coincidence or the
presence of a certain third, unseen factor (referred to as a confounding factor or lurking variable).
Suppose there is found to be a correlation between A and B. Aside from coincidence, there are three
possible relationships:
A causes B,
B causes A,
OR
C causes both A and B.
In the last case there is a spurious correlation between A and B. In a regression model where A is regressed
on B but C is actually the true casual factor for A, this misleading choice of independent variable (B instead
of C) is called specification error.
Scatter Diagram
Suppose we are given two paired variable values - one is related to the heights of the students while the
other is related to the weights of the students. They are plotted on the coordinate plane using x-axis to
represent heights and y-axis to represent weights. Then each paired observation shall have one point on the
graph. The graphical representation of dots so obtained is called Scatter Diagram. The different graphs
shown below illustrate the different type of correlations.
.
Y
High degree positive correlation
High degree negative Correlation
68
P er f ect p os itive cor relatio n
Perfect neg ative correlatio n
X
M o der ate d egr ee -ve cor r elatio n
Mo der ate degr ee + ve correlatio n
L o w d egree - ve corr elation
X
L ow deg ree +v e corr elation
X
No correlation
69
3.3
Correlation Coefficient
The Coefficient of Correlation r is a measure of the degree of linear relationship between two variables,
say, x and y i.e. it measures the degree of association between the two values of related variables given in
the data set.
(i) The coefficient of correlation r (also known as Karl Pearsons Coefficient of Correlation) is given
by:
r=
(ii) If the mean of variables x and y, are denoted by and , than we can apply the following formula
to calculate the coefficient of correlation r:
difference of values from
its arithmetic mean
r=
(iii)If the difference from any constants A or B (other than arithmetic mean) is taken the following
formula can be used to calculate the coefficient of correlation r.
r=
The value of coefficient of correlation r lies between - 1 and + 1. If for a given two sets or data:
r = +1, the two sets or data are said to be perfect positively correlated.
r = - 1, the two sets or data are said to be perfect negatively correlated.
r = 0, the two sets or data are said to be uncorrelated, that is, there is absence of any linear
relationship between the variables. However, there may exist some other form of relationship between
them.
r = 0.8, the two sets or data are said to be strongly correlated.
r = 0.2, the two sets or data are said to be weakly correlated.
Properties of Coefficient of Correlation r

1. The value of r is independent of the origin of reference and the scale of reference i.e. r is not
affected by addition or subtraction of a constant to the values of either or both variables. It is also,
unaffected by the multiplication or division of the values of either or both variables, by a constant.
2. The value of r is free from any of the units. In other words, r will have a definite meaning,
whether the units of x and y variables are comparable or not.
3. The coefficient of correlation is symmetrical in two variables.
4. The value of r lies between + 1 and - 1.
70
Example1: Calculate the coefficient of correlation between X and Y series:

X:
Y:
1
8
3
12
5
15
7
17
8
18
10
20
Solution:
Table for Calculation
X
X2
Y2
XY
1
3
5
7
8
10
8
12
15
17
18
20
1
9
25
49
64
100
64
144
225
289
324
400
8
36
75
119
144
200
X = 34
Y = 90
X2 = 248
Y2 = 1446
XY = 582
The coefficient of correlation r is given by:

r=
- - - ( 1)
Substituting the values from the table in the equation ( 1 ), we obtain:
r=
r=
r=
= 0.988
Example 2: Calculate the coefficient of correlation between the height of father and son from the following
data:
Height of Father (in Inches) (x) :
Height of Son (in Inches) (x) :
64
66
65
67
66
65
67
68
68
70
69
68
70
72
Solution:
Height of
Father (in
inches)
x
64
65
66
67
68
69
70
Height of
Son (in
inches)
y
66
67
65
68
70
68
72
x = 469
y = 476
dx
(x 67)
dy
(y 68)
dx2
dy2
dx.dy
-3
-2
-1
0
+1
+2
+3
-2
-1
-3
0
+2
0
+4
9
4
1
0
1
4
9
4
1
9
0
4
0
16
+6
+2
+3
0
+2
0
+12
dx = 0
dy = 0
dx2 = 28
dy2 = 34
dx.dy = 25
71
= 67
= 68
=
=
The coefficient of correlation 'r' is given by:

r=
.
2.
- - - ( 1)
dy 2
Substituting the values from the Table in equation (1), we obtain

r=
25
28 . 43
25
r = 30.85 = 0.81
Spearmans Rank Correlation
Ranked data are defined as data which are arranged in numerical order, usually from largest to smallest and
numbered 1, 2, 3, . . . The correlation of such type of data is known as Rank Correlation, as here only ranks
are considered and denoted by R.
The coefficient of Rank correlation R is given by:
R = 1-
where, n is the total number of individuals and d is the difference in ranks.
where, n is the total number of individuals and d is the difference in ranks.

Properties of Rank Correlation Coefficient:
Rank Correlation Coefficient lies between +1 and - 1, including both the values.
If dj stands for the difference in the ranks of the jth individual and if dj = 0 for all individuals, then
Rank correlation coefficient R = 1.
Example 3: The marks obtained by 8 students in Physics and Chemistry are given in the following Table:
Marks in Physics :
52
54
67
82
98
90
69
76
Marks in Chemistry :
11
23
36
56
37
12
25
Calculate the Rank Correlation Coefficient.

Solution:
First change the given values into ranks for each series then after giving ranks to the students in Physics and
Chemistry and calculating the difference d in ranks, we obtain the values shown in following Table:
72
Marks in
Physics
(x)
52
54
67
82
98
90
69
76
Marks in
Chemistry
(y)
11
7
23
36
56
37
12
25
Rank in
Physics
Rank in
Chemistry
Difference in
Rank (d)
d2
8
7
6
3
1
2
5
4
7
8
5
3
1
2
6
4
+1
-1
+1
0
0
0
-1
0
1
1
1
0
0
0
1
0
d2 = 4
The Rank Correlation Coefficient is given by:

R = 1-
)
6. 4
R = 1- 8 (64 1)
24
= 1- 504
= 1 - 0.04761
= 0.95238
Example 4: Ten competitors in a beauty contest are ranked by three judges in the following order:
1st judge
10
2nd judge
10
3rd judge
10
Use the Rank Correlation Coefficient to determine which pair of judges has the nearest approach to
appreciation of beauty.
Solution:
In order to find out which pair of judges has the nearest approach to appreciation of beauty, we compare
Rank Correlation between the judgments of:
(i) 1st judge and 2nd judge
(ii) 2nd judge and 3rd judge
(iii) 1st judge and 3rd judge
73
Rank by 1st
judge
R1
Rank by 2nd
judge
R2
Rank by 3rd
judge
R3
(R1 R2)2
(R2 R3)2
(R1 R3)2
1
6
5
10
3
2
4
9
7
8
3
5
8
4
7
10
2
1
6
9
6
4
9
8
1
2
3
10
5
7
4
1
9
36
16
64
4
64
1
1
9
1
1
16
36
64
1
81
1
4
25
4
16
4
4
0
1
1
4
1
N = 60
N = 60
N = 60
d2 = 200
d2 = 214
d2 = 60

R = 1-
The Rank Correlation Coefficient between the judgments of 1st and 2nd judges::
6 . 200
R (I and II) = 1- 10 (100 1)

= 1-
1200
990
= 1 - 1.212
= - 0.212
The Rank Correlation Coefficient between the judgments of 2nd and 3rd judges::
6 . 214
R (II and III) = 1- 10 (100 1)

= 1 - 1.297
= - 0.297
The Rank Correlation Coefficient between the judgments of 1st and 3rd judges::
6 . 60
R (I and III) = 1- 10 (100 1)

360
= 1- 990
= 1 0.3636
= 0.636
Since coefficient of correlation is maximum in the judgments of the first and third judges we conclude that
they have the nearest approach to appreciation of beauty.
Equal Ranks If a tie occurs in the ranks of two individuals, they are assigned the average of ranks that they
would have obtained. For example, if two individuals are tied up at the fifth rank, each may be ranked ( 5 +
74
6) / 2 = 5.5 and next is then ranked 7. Where equal ranks are assigned to some entries an adjustment in the
above formula for calculating the rank coefficient of correlation is made.
The adjustment consists of adding (m3 m) to the value of d2, where m stands for the number of items
whose ranks are common. If there are more than one such group of items whose ranks are common, this
value is added as many times the number of such groups.
Where equal ranks are assigned to some entries the coefficient of Rank correlation,
R is given by:
R = 1-
Example 5: Obtain the rank correlation coefficient between the variables X and Y from the
following pairs of observed values:
X:
50
55
65
50
55
60
50
65
70
75
Y:
110
110
115
125
140
115
130
120
115
160
Solution:
X
Rank X
R1
Rank Y
R2
(R1 R2)2
d2
50
55
65
50
55
60
50
65
70
75
2
4.5
7.5
2
4.5
6
2
7.5
9
10
110
110
115
125
140
115
130
120
115
160
1.5
1.5
4
7
9
4
8
6
4
10
0.25
9
12.25
25
20.25
4
36
2.25
25
0
d2 = 134
In series X = 50 is repeated thrice (m = 3); 55 has been repeated twice (m = 2), 65 has been
repeated twice (m = 2). In series Y; 110 has been repeated twice (m = 2) and 115 thrice (m =
3)
R=1-
R = 1-
= 1-
= 1-
6 134+
1
12
33 3 +
1
12
23 2 +
1
12
1
12
23 2 +
10(10 2 1)
6 {134+ 2 + 0.5+ 0.5+0.5+2}

990
6 (139.5)
990
= 1-
837
990
= 1- .845 = 0.155
75
1
12
23 2 +
33 3
3.4
Simple Linear Regression
Sir Francis Galton introduced the term Regression during the end of the 19th century. In todays world the
term Regression is used to measure the average relationship between the correlated variables. In other
words, Regression is the prediction of unknown values of one variable from the known values of other
variable.
Regression Analysis is an important technique used to analyze the relationship between a dependent variable
and one or more independent variables. More specifically, regression analysis helps one to understand how
the typical value of the dependent variable changes when any one of the independent variables is varied,
while the other independent variables are held fixed. It helps to estimate the average value of the dependent
variable from given values of independent variables. It is one of the most important statistical tools which is
extensively used in almost all sciences - Natural, Social and Physical. It is specially used in business and
economics to study the relationship between two or more variables that are related causally e.g. demand
and supply, cost functions, production and consumption functions and so on.
Linear Regression involves a relationship between two variables only. There is a linear regression between
the variables under study, if the regression curve is a polynomial of degree one. In other word, it means that
if the set of observations for these two variables are plotted, a straight line is drawn through the scattergram.
The linear relationship between two variables x and y is represented by the following equation, which is a
polynomial of degree one.
y = a + bx,
where, a and b are the constant determining the position of the line.
The slope of the line is denoted by b and intercept on the y-axis is given by a. This relationship is called
Linear Regression of y on x. Here b is called regression coefficient of y on x. Similarly, the Linear Regression
of x on y is given by equation:
x = c + dy,
where, c and d are constants. The constant d is called regression coefficient of x on y and the constant c is
the intercept on x- axis.
It may be that all the points do not lie on the straight line equation which represents the linear relationship
between two variables. The straight line equation shows the average relationship between the values of the
two variables.
3.5
Estimating Linear Regression
The estimation or prediction of future production, consumption, prices, investments, sales, profits, income
etc. are of very great Importance to business professionals. Similarly, population estimates and population
projections, GNP, Revenue and Expenditure etc. are indispensable for economists and efficient planning of
an economy. Regression estimates the changes in the value of dependent variables with respect to some
values of independent variable. These estimation are done by the help of regression lines which arepolynomials
of degree one, having two variables.
Equation of Regression Lines
The straight line depicting the relationship between two variables is called Line of Regression or Regression
Line. A Line of Regression is the one which gives the best fit using Principal of Least Square.
76
In order to obtain the equation of regression line, we will be using the Method of Least Squares. By the help
of this method, we will obtain the best fitting line to the set of points on the scatter diagram. Suppose, the
equation of line of best fit of x, be
y = a + bx
where, a and b are constants.
Let X and Y be the deviation from the respective means
X=xWhere ,
and
and
i.e.
, Y= y-
are mean of x- series and y-series, respectively.
Let Ai be any point in graph and Bi be the corresponding point on the line (see the figure). Then, the distance
between these points is represented by Ai - Bi .
Y
Ai
Bj
Bi
Aj
If S be the sum of squares of such distances, then
S = (Ai - Bi)2
= (Yi a -bXi)2
The principal of Least Squares says that sum of the squares of such distances be minimized. For this we
have to choose those values of a and b for which this sum S is minimum. Each pair of a and b
represents one regression line passing through the scattergram. We need to find that line for which this sum
of squares is minimum.
Equation of line of best fit for Y on X is given by:
(y-
)= r
(x-
Similarly, we can obtain the equation of regression line of X on Y,

(xbyx = r
and bxy = r
)= r
(y-
are known as regression coefficient of y on x and x on y,
respectively.
77
There are other variations of the above formula, which can also be used to find the regression equations,
these are:
Solution of Normal Equations:
The regression equation of X on Y is expressed as:
X = a + b.Y
To determine the values of a and b, the following two normal equations are to be solved
simultaneously:
X = N. a + b . Y
XY = a . Y + b . Y2
Solving these equations, values of a and b are obtained as follows:
X b. Y
=
where,
N. XY X . Y
N. Y2 (Y)2
The regression equation of Y on X is expressed as:

Y = c + d.X
To determine the values of c and d, the following two normal equations are to be solved
simultaneously:
Y = N. c + d . X
XY = c. X + d . X2
Similarly solving these equations, values of c and d are obtained as follows:
Y d X
=
where,
N. XY X . Y
N. X2 (X)2
Example 6: Marks obtained by a student in Physics and Maths (out of 100) are given in the following
Table:
Physics (x)
80
45
55
56
58
60
65
68
70
75
85
Maths (y)
82
56
50
48
60
62
64
65
70
74
90
Find the equations of lines of regression.

Solution:
Take, X = x - 65 and Y = y 70
78
Physics
(x)
Maths
(y)
X = x - 65
Y = y - 70
X2
Y2
XY
80
45
55
56
58
60
65
68
70
75
85
82
56
50
48
60
62
64
65
70
74
90
15
-20
-10
-9
-7
-5
0
3
5
10
20
12
-14
-20
-22
-10
-8
-6
-5
0
4
20
144
196
400
484
100
64
36
25
0
16
400
x = 717
y = 721
X = 2
Y = -49
225
400
100
81
49
25
0
9
25
100
400
X2 =
1414
180
280
200
198
70
40
0
-15
0
40
400
XY =
1393
First we will calculate the standard deviation of x and y i.e.

correlation 'r'.
Y2 = 1865
and
and coefficient of
Substituting the values from the Calculation Table in the formulas for coefficient of
correlation and standard deviation, we obtain,
Thus,
Now as regression coefficient of x on y is bxy and y on x is byx we have

bxy = r
byx = r
(0.92). (11.34)
12.24
(0.92). (12.24)
12.34
If the mean of x-series and y-series are

X = 65 +
Y = 70 +
= 0.85
= 0.99
and , respectively, then
= 65 + 11 = 65.2
= 70 +
(49)
11
= 65.55
79
y-
=r
(x -
Hence, the equation of regression line of y on x is:

y - 65.55 = (0.99) . (x 65.2)
y = (0.99) . x + 1.002
Also, the equation of regression line of y on x is:
x-
=r
(y -
Therefore, the equation of regression line of x on y is:

x - 65.2 = (0.85) . (y 65.55)
x = (0.85) . y + 9.48
Example 7: For certain X and Y series which are correlated, the two lines of regression are:
5X 6Y + 90 = 0
15X 8Y 130 = 0
Find the means of the two series and the correlation coefficient.
Solution:
(i) Finding mean of the two series:
5X 6Y = 90
15X 8Y = 130
Multiplying equation (i) by 3, we get
15X 18Y = -270
15X 8Y = 130
10Y = 400
Y = 40
Putting the value of Y in eq. (i), we get X
5X 6 (4) = -90
5X = 90 + 240
5X = 150
X = 30
(ii) Finding correlation coefficient:
Let us assume that equation (i) is the regression equation of X on Y;
5X = 6Y 90
6
X = Y 18
bxy =
5
6
5
Taking eq. (ii) as the eq. of Y on X,

Y = 15X + 130
8Y = 15X 130
Y=
byx =
15
8
130
8
15
8
80
. . .(i)
. . .(ii)
Since both the regression coefficients are exceeding one, our assumption is wrong. Hence, eq. (i) is the
regression eq. of Y on X:
6 Y = 5X 90
6Y = 5X + 90
6
Y=
byx =
X + 15
5
5
6
Eq. (ii) is the regression eq. of X on Y;

15 X = 130 + 8Y
130
X=
bxy =
15
15
8
15
r=
=
15
= 0.667
Example 8: Height of fathers and sons are given below. Find the height of the son when the height of the
father is 70 inches.
Height of Father
71
68
66
67
70
71
70
73
72
65
66
69
64
65
63
65
62
65
64
66
59
62
(inches)
Height of Son
(inches)
Solution:
To find the height of son when the height of father is given we have to fit the regression of
son (y) on father (x).
Father
(x)
X=x-
X2
Son
(y)
Y=y-
Y2
XY
71
68
66
67
70
71
70
73
72
65
66
+2
1
3
2
+1
+2
+1
+4
+3
4
3
4
1
9
4
1
4
1
16
9
16
9
69
64
65
63
65
62
65
64
66
59
62
+5
0
+1
1
+1
2
+1
0
+2
5
2
25
0
1
1
1
4
1
0
4
25
4
+10
0
3
+2
+1
4
+1
0
+6
+20
+6
x = 759
X = 0
X2 = 74
y = 704
Y = 0
Y2 = 66
XY = 39
81
x=
y=
r
704
11
759
XY
11
=64
=69
39
= X 2 = 74 =0.527
We know that the equation of regression line of y on x is given by:

y-
=r
(x -
y 64 = 0.527 (X 69)
y 64 = 0.527 X 36.36
y = 0.527X + 27.64
For x= 70, y is given by:
y = 0.527 (70) + 27.64
= 64.53
Hence, the height of son when the height of father is 70 inches shall be 64.53 inches.
Properties of Regression Coefficient
1. The Coefficient of Correlation is the Geometric Mean of the two coefficient of Regressions.
2. Both the coefficients must be of same sign i.e. either both are positive or both are negative, but one
negative and other positive is not possible.
3. Arithmetic Mean of coefficient of regression is greater than the coefficient of correlation.
4. The coefficient of regression is independent of change of origin but not of scale.
5. The correlation coefficient and the two regression coefficients have same sign.
6. Both the coefficient of regressions cannot be greater than one and product cannot be greater than
one i.e. bxy . byx 1. But, one of them can be greater than one, individually; provided the above
condition is satisfied.
3.6
Difference between Correlation and Regression
Following are the differences between correlation and regression.
Correlation coefficient is a symmetrical function between x and y but the regression coefficient are
not symmetrical functions between x and y.
Correlation coefficient is independent of change of scale and origin, whereas, the regression coefficients
are independent of the origin but not of scale.
Correlation coefficient is a relative measure of relationship, whereas, regression coefficient is an

absolute measure of relationship.
Correlation coefficient is merely a tool of ascertaining the degree of relationship between two variables
whereas in regression analysis one variable is taken as dependent while the other as independent; to
estimate the value of one from another.
82
3.7
Summary
Correlation and Regression are important statistical tools for measuring the degree of association between
series of pairs of observations of two or more variables. These tools are used for studying and measuring
the extent of relationship between two or more variables.
Correlation is the possible relationship between two series of observations such that the changes in the
values of one series are accompanied by changes in the values of other series. Two variables are said to be
positively correlated if an increase (or decrease) in the value of one variable is accompanied by an increase
(or decrease) in the value of other variable. Two variables are said to be negatively correlated if the increase
(or decrease) in the value of one variable is accompanied by decrease (or increase) in the value of other
variable. Coefficient of Correlation shows the strength (magnitude) and direction (direct or indirect) of the
relationship between two variables. It stands for relative measure of relationship between two variables.
Regression is used to measure the average relationship between the correlated variables. It uses the known
value of one variable to predict or estimate the unknown value of the other variable. Regression stands for
absolute measure of relationship. In Linear Regression, only two variables are involved in a relationship.
The Method of Least Squares was used to find the best regression lines, which gives best fit for the set of
observations. By the help of regression lines, we are able to estimate the value of x by the value of y and
vice versa.
3.8
Key Words
Bivariate Distribution - a distribution in which each individual of the set can generate two values.
Scatter Diagram - a diagram of plotted points that show relationship between two sets of data.
Linear Correlation- Correlation between two variables in which relationship is in the form of
polynomial of one degree.
Coefficient of Correlation - measure the intensity of relationship between two variables.
Coefficient of Regression - indicate relation of average of one variable with respect to the average
of other variable (Assumed to be independent).
3.9
1.
Calculate the coefficient of correlation for the following observations:

X
2
(Ans: 0.47)
2.
Calculate the coefficient of correlation between the values of x and y from the following:
X
10
12
15
17
18
20
(Ans: 0.98)
83
3.
Calculate the coefficient of correlation from the following data:

X
100
200
300
400
500
600
700
30
50
60
80
100
110
130
(Ans: 0.997)
4.
Eight student obtained the following marks (out of 100) in Maths and Chemistry.
Marks in Maths
15
13
27
45
20
60
20
75
Marks in Chemistry
50
30
55
25
30
10
30
70
Calculate the rank correlation coefficient.

(Ans: 0.02)
5.
The marks secured by recruits in the selection test (X) and in the proficiency test (Y) are given
below:
X
10
15
12
17
13
16
24
14
22
30
42
45
46
33
34
40
35
39
Calculate the rank correlation coefficient

(Ans: 0.4)
6.
The coefficient of rank correlation of marks obtained by 10 students in statistics and accountancy
was found to be 0.2. It was later discovered that the difference in ranks in two subjects obtained by
one of the students was wrongly taken as 9 instead of 7. Find the correct coefficient of rank
correlation.
(Ans: 0.394)
7.
Heights of fathers and sons (in inches) are given in the following table:
Heights of father
65
66
67
67
68
69
71
73
Heights of Son
67
68
64
68
72
70
69
70
Obtain the two lines of regression and calculate the expected average height of the son when the
height of father is 67.5 inches.
(Ans: y = 0.421x + 39.29; x = 0.524y + 39.29; height of son is 68.19)
8.
For 50 students of a class the regression equation of marks in Statistics (X) on the marks in
Accountancy (Y) is 3 Y 5 X + 180 = 0. The mean marks in accountancy is 44 and variance of
marks in Statistics is 9/16th of the variance of marks in Accountancy. Find the mean marks in
Statistics and the coefficient of correlation between marks in two subjects.
(Ans: + 0.8)
9.
The following table shows the ages (X) and blood pressure (Y) of 8 persons:
X
52
63
45
36
72
65
46
25
62
53
51
25
79
43
60
33
84
Obtain the regression equation of Y on X and find the expected blood pressure of a person who is
49 years old.
(Ans: 49.502)
10.
From the following data given below find:

(a)
The two regression equations
(b)
The coefficient of correlation between the marks in Economics and Statistics.
(c)
The most likely marks in Statistics when the marks in Economics are 30.
Marks in Economics
25
28
35
32
31
36
29
38
34
32
Marks in Statistics
43
46
49
41
36
32
31
30
33
39
(Ans: (a) Regression eq. X on Y: X = 40.892 0.234Y; Regression eq. Y on X: Y = 59.248 0.664X;
0.394; 39)
3.11 References
Introduction to Mathematical Statistics, P.G. Hoel, John Wiley and Sons.
Methods of Correlation Analysis, Ezekiel, John Wiley and Sons.
Mathematical Statistics, O.P. Gupta, Kedarnath Ramnath and Co.
85
Unit - 4 : Business Forecasting and Time Series

Unit Structure:
4.0
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
4.10
4.11
4.12
Objectives
Introduction
Objectives and Importance
Methods of Business Forecasting
Time Series
Importance and Utility
Components of Time Series
Preliminary Adjustments for Time Series
Methods of Finding Trends
Summary
Key Words
Self Assessment test
References
4.0
Objectives
After studying this unit, you should be able to understand :

The concept of business forecasting
The objectives and importance of forecasting in business
The concept of Time Series Analysis
Utility of Time Series
The various components and analysis of time series
The measurement of trend
4.1
Introduction
Forecasting refers to the estimation of future based on the data related to past years. Forecasting may be
done related to various areas of economy, business and other allied areas.
Forecasting is more important for business decision making. The success of business depends upon the
prompt decision making based on appropriate data collection, in depth analysis and interpretation. Analysis
is generally related to the past data and finding out trends for estimating future movements of demand, sales,
cost and other concerned areas of business.
Business forecasting refers to the statistical analysis of the past and current movements in a given time
series, so as to obtain clues about the future pattern of the movements.
Neter and Wasserman
Following features may be considered from above explanation:
1. It is based on past data.
2. Business forecasting may be done for short period or long period.
86
3. Business forecasting may be related to particular variable or multiple variables.

4. Business forecasting can be useful in business analysis and decision making.
4.2
Objectives and Importance
Objectives of Forecasting in Business

Forecasting may become an important tool in business. In todays era, entrepreneurship is promoted by
government and various subsidies are planned as well as provided for initiation. Thus, starting up a new
venture or new business has became an easier task. But initiation alone is not enough, the real success of the
business depends upon how well the entrepreneur uses his following skills for business forecasting:
Conceptual Skills: It refers to the basic understanding of the operating activities of the business. It
helps to identify the critical success factors for the business and also help to locate those variables
which need proper attention and apt forecasting.
Diagnostic Skills: It requires appropriate method of analyzing the data collected. If an entrepreneur
lacks these skills then he may not find out the root cause of the problem and may not be able to take
appropriate decision.
The above mentioned skills will promote approximate forecasting which would further decide the future
success and failure of the business. Forecasting helps in minimizing the uncertainties related to future events.
Forecasting also acts as a foundation for decision making related to capital investment, price change,
financing source, cost of capital, innovations in production techniques, markets, product design etc. It can
be safely assumed that accuracy in forecasting cannot be always 100%. This is because the future is uncertain
and uncertainty may be reduced but cannot be eliminated completely. Thus, a margin of error is allowed
which is referred to as probable error.
Importance of Business Forecasting:
Business forecasting assumes great relevance when its results are implemented in various functions of
management:
Planning: The planning is based on various assumptions about future conditions. Forecasting helps
in estimating the future conditions and thus helps in revising the planning premises. It provides
stability of plans for a longer span of time.
Better Information System: Forecasting is useful in gaining information about diverse fields which
can be used in strategy formulation, planning marketing activities, formulating financial policies and
improving the operating efficiency.
Improving productivity and profitability: By analyzing the forthcoming events, productivity and
profitability can be improved by eliminating waste such as:
Plant shut-downs
Lost sales
Lost customers
Expensive expediting
87
Missed strategic opportunities.

Financial and Strategic Importance: Forecasted data may be used in better strategy formulation
as marketing, production, demand figures may be extrapolated. Financial policies can be formulated
based on forecasted data which may result in better allocation of resources, increased profitability
and higher return on investment.
4.3
Methods of Business Forecasting
Business forecasting may be done by various methods. As per the need of the business, appropriate method
should be selected. Following important methods are discussed below:
Main Methods of Forecasting:
1. Time Series Analysis:
Time series analysis may be used for the purpose of business forecasting. Forecasting through this method
can be done only when a series of data for several years is available. This method assumes that the business
data reflects a trend which may be long term, cyclical or seasonal in nature. This trend may be identified by
using this method and then future trend is predicted on its basis. Time series analysis is based upon additive
or multiplicative model. (Discussed in detail in the later half of this unit)
Merits
i. It is an easy method to apply if appropriate data series is available.
ii. As the future is forecasted on the basis of past business data, the forecasting will be more reliable.
Demerits
i. In this method, the forecasting is based on guess work not on a scientific method because the past and
present conditions are rarely found to be similar.
ii. It is very difficult to select the past period with the same business conditions like present.
2. Survey Method:
Survey means to approach the respondents comprehensively or selectively. Survey may be of two types:
Census survey and Sample survey. Census survey means to approach each person under the study to
collect the required information. It consumes more time and money. On the other hand, sample survey may
be conducted in which a sample of few persons is selected from the population under study randomly or
according to the judgment of the researcher. Sample survey is convenient for data collection and consumes
less time and money.
Field surveys can be conducted to gather information about the intentions and attitudes of the concerned
people for the future. Such surveys may be conducted amongst consumers about the likely expenditures on,
say consumer durables. Both, quantitative and qualitative information may be gathered.
The survey may be conducted by using questionnaires, interviews or schedules. The data collected is then
tabulated for analysis. On the basis of analysis future trend may be predicted.
3. Opinion Poll:
This method is used when valuable information is assumed to be obtained not from the customer but from
others who are related closely to the subject under study. This method is useful in behavioral or marketing
data collection for future forecasting.
88
Opinion poll is the survey of opinion of knowledgeable persons or experts in the field whose views carry lot
of weightage in the area of study. For example, a survey of opinion of the sales representatives, wholesalers
or marketing experts may be helpful in gathering data about the consumers intention or expectations and
then on its basis demand projections may be made.
This approach is the vicarious approach i.e. using others experience to gather knowledge. This method is
reliable as the forecasting based on this method has a greater tendency of accuracy.
4. Exponential Smoothing
This method reduces the extremeness of past values and gives more importance to the current values.
Exponential smoothing is special kind of weighted averages which is more useful in short-term forecasting.
This method gives less weights to the past values and higher weights to the current values. The effect of past
data is automatically reduced as more weightage is given to recent year values. This method is more useful
and gives reliable results on which effective business decisions may be based. Thus, this method is regarded
as the best method of business forecasting as compared to other methods.
Merits
i. Forecasting is made on the basis of past conditions so it is more reliable.
ii. This method is helpful in short term forecasting.
iii. This method may be conveniently applied on computer.
Demerits
i. It cannot be successfully applied for long term forecasting.
ii. Present conditions are given more weightage than the past conditions.
iii. This method involves heavy calculations as it involves calculation of moving average and then applying
weights for further calculations.
5. Regression Analysis:
It is a statistical technique used to find the nature of relationship between variables. The identification of
relationship can be used for forecasting of either of the variables. There may be two or more than two
variables under study. For more than two variables, multiple regression is used.
Under this method, first the variables are identified as dependent and independent variables. Then, the value
of dependent variable may be forecasted by identifying its nature of relationship with the independent variable.
The relationship is expressed in two forms of equations stated as follows:
i)
X on Y : The value of dependent variable X may be worked out when independent variable Y is
known.
ii)
Y on X : The value of dependent variable Y (assuming it as dependent) may be worked out when
values of variable X are known and assumed as independent.
6. Barometric Techniques:
Barometer means which indicates about something. A barometric technique involves the usage of economic
barometers or business barometers for forecasting purpose. Business barometers are those factors which
represent the change in the economic conditions of the economy and thus affect the individual business unit
as a whole. A study of these barometers can be used in predicting the future trends in business conditions.
89
Some of the economic indicators are Gross Domestic Product, National Income, Consumer Price Index
etc.
The business barometers may be classified into three categories:
i) General Index of Business Activity: These barometers are related to general business activities.
A separate index for every type of unique business activity is constructed which help in the formulation
of the economic policies of the country.
ii) Barometers for Specific Business Activities: These indices are worked for specific
infrastructural business activities which act as a supplementary index to the above indices. E.g. Iron
& steel, petrochemical business etc.
iii) Barometers for individual business firm: These indices are calculated in relation to specific
business firm and may be used in estimating future variations in the business activities.
7. Modern Econometric Method:
This method is purely mathematical and statistical based. It involves many sets of simultaneous equations to
be prepared. Solving these equations manually is very difficult, so the usage of advanced technology is a
must. Computer softwares and advanced computing programmes are now a day very popularly used; using
this method for business forecasting became easy.
Merits:
i. Accurate and reliable results are obtained under this method because it is a scientific method where
computer is used for calculations.
ii. This method is more logical as it explains the interrelationship between various economic variables.
Demerits:
i. This method is difficult and complicated as it involves a number of equations to be formulated and then
solving them.
ii. If appropriate data series is not available, it will make this method unusable.
iii. It is based on growth model which is difficult to build up.
Activity A
1.
Identify the problems pertaining to selection of the method for forecasting sales in any organization
familiar to you and choose the most appropriate method.
4.4
Time Series
A time series may be defined as a sequence of repeated measurements of a variable made periodically
through time.
Cecil H. Mayers
A time series may be defined as a sequence of values of some variable corresponding to successive points
in time.
W.Z.Hersch
Time series is a collection of readings belonging to different time periods in the order of their occurrences.
Following are the examples of time series:
90
1.
Monthly consumption.
2.
Annual Production.
The following is the time series data which can be further used for forecasting:
Year
2006
2007
2008
2009
2010
2011
4.5
Sales (in lakh)

10
12
16
8
6
14
Importance and Utility
Importance or Utility of Time Series Analysis:

Based on analysis of past behavior: - It is assumed that the future trends are based on the
historical trends. Thus, such analysis will be extremely helpful in predicting the future behavior.
Creating a road map for the future: It helps in planning future operations. Business operations
can be run more efficiently and effectively by predicting the future.
It facilitates comparison: - Different time series are often compared and important conclusions
drawn there from. The trends reflected may be different like one time series may have seasonal
variations and the other time series may have more of irregular fluctuations component.
It helps in planning future operations: - Plans for the future can not be made based on the result
of time series. Statistical techniques have been evolved which enable time series to be analysed in
such a way that the influences which have determined the form of that series may be ascertained. If
the regularity of occurrence of any feature over a sufficient long period could be clearly established
then, within limits, prediction of probable future variations would become possible.
It helps in evaluating current actions: - The actual performance can be compared with the
expected performance and the cause of variation analysed. For example, if expected sale for 201011 was 1,000 laptops and the actual sale was only 700, one can investigate the cause for the
shortfall. Time series analysis helps to analyse one variable at a time as it assumes other things
constant. For example, if we know how much is the effect of seasonality on business we may find
ways and means of ruling out the seasonal influence or decreasing it by producing commodities with
complementary seasons.
4.6
Components of Time Series
Time series analyses many variations in the available data. The variations may be of different forms. These
diverse changes can be categorized under following four categories, referred to as Components of Time
Series.
1.
Secular Trend
2.
Cyclical Trend
3.
Seasonal Variations
4.
Irregular Variations
91
Components
Long-Term
Secular Trend
Short-Term
Cyclical
Seasonal
Irregular or
Erratic
Regular
The components of time series as depicted by the above diagram represent fluctuations over time in various
business activities which may be long term or short term.
Long Term: The fluctuations which are long term in nature may be worked out in two waysSecular trend and Cyclical trend.
Short Term: The fluctuations of short term nature may be represented seasonally or in irregular
formats.
1. Secular Trend: General movement persisting over a long period of time (several time periods) is called secular trend.
Secular trend movements are due to the factors such as population change, technological progress and
large scale shifts in consumer tastes.
Features:
It involves long term steady trends created by Growth factors or Decline factors.
They represent slow and continuous trends.
The short term happenings do not affect this trend.
Such trends may be notified widely in the economic conditions of the country.
2. Cyclical Movements or Cyclical Variations: Cyclical movements refer to the long term (over several time periods / years) movements which are consistent
and generally periodic in nature.The values increase over some years reach peak and start downward
movement and touch a bottom; again start increasing and reach a top; this complets one business cycle. The
time period of business cycles varies and so does the top and bottom levels. These movements are important
in analysis of business activities as business cycle represents cyclical movements of ups and downs.
Features:
They depict the turning points of the business activities which further help in systematic future planning.
Normal range of these variations is from 2 to 15 years.
It analyzes the expansion and contraction phase of the economic activities.
92
3. Seasonal Movements or Seasonal Variations: These refer to the identical patterns that a time series appears to follow during corresponding months or
quarters of successive years. The type of fluctuation, which completes the whole sequence of changes
within the span of a year and has the same pattern year after year, is called seasonal variations.
Features:
The seasonal movements are short term in nature and are repeated perpetually every year.
The major factors causing these fluctuations are:
a)
Climate and weather
b)
Customs and traditions
c)
Habits
d)
Fashion
An analysis of these variations will help to tap the unused resources and their optimum utilization.
4. Irregular or Random Movements: The variations which do not follow a defined pattern and their reoccurrence is not certain, nor may be
depicted is termed as Irregular Variations. These fluctuations occur generally due to contingent happenings
or unexpected situations.
Features:
They cannot be predicted as they do not repeat in a definite pattern.
They are caused by earthquakes, floods etc. and very rapid technological changes.
These variations may act as a turning point of the business which may be taken as a threat or an
opportunity.
Time Series Models
An analysis of time series may lead to a combination of the above four variations and so each variation must
be identified to work out a trend value for referring future values.
There are two types of time series models:
1. Additive Model : It assumes that the original data is a sum of the four variations.
Y=T+S+C+I
Y = Original Data
T = Trend value
S = Seasonal Variations
C = Cyclical Variations
I = Irregular Variations
2. Multiplicative Model or Classical Model: It assumes that the original data is a product of the four
components. This model is widely used.
93
Y=TxSxCxI
= TSCI
4.7
Preliminary Adjustments for Time Series
Preliminary adjustments refers to those changes required to eliminate the effect of abnormal events or
certain peculiarities which may effect the time series data. These adjustments are explained below:
1. Calendar Adjustments: - To eliminate certain differences which are caused by peculiarities of our
calendar like no. of days in month, holidays, festivals etc.
Most of the time series are available in a monthly form and it is necessary to recognize that the month is a
variable time unit. The actual length of the shortest month is about 10% less than that of the longest, and if
we take into account holidays and weekends, the variation may be even greater. When we study the
production level or sales on a monthly basis, identifiable variations may be observed in different months.
Thus the purpose of adjusting for calendar variation is to eliminate certain counterfeit differences which are
caused by peculiarities of our calendar. The adjustment for calendar variation is made by dividing each
monthly total by the number of days in the month.
Daily average for each month = Monthly Total / No. of working days
Comparable (adjusted) monthly data may then be obtained by multiplying each of these values by 30.4167,
the average number of days in a month. (In a leap year this factor is 30.5)
2. Population Change Adjustments: Certain types of data call for adjustment for population changes. Changes in the size of population can easily
distort comparisons of income, production and consumption figures. For example, national income may be
increasing year after year, but per capita income may be declining because of greater pressure of population.
In such cases where it is necessary to adjust data for population changes, a very simple procedure is
followed, i.e., the data are expressed on a per capita basis by dividing the original figures by the appropriate
population totals.
3. Adjustment for price changes: - Adjustment for price rise and fall should be made.
An adjustment for price changes is necessary whenever we have a value series and are interested in quality
changes alone. Because of rising prices the total sales proceeds may go up even when there is a fall in the
number of units sold. For example, if in 2010, 1,000 units of a commodity that is priced Rs. 20 are sold, the
total sale proceeds would be 1,000 x 20 = Rs. 20,000. Now suppose in 2011 the price of the commodity
increases from Rs. 20 to Rs. 22. If the sales do not decline, the total sales proceeds will be 1,000 x 22 = Rs.
22,000. This increase in sales proceeds, i.e., Rs. 2,000, is not due to increase in the demand but due to the
rise in price from Rs. 20 to Rs. 22.
The effect of price changes can be eliminated by dividing each item in a value series by an approximate price
index.
4. Comparability: - For any meaningful analysis of time series, it is necessary to see that the data are
strictly comparable throughout the time period under investigation.
For example, if we are observing a phenomenon over the last 20 years, the comparability may be observed
by differences in definition, differences in geographical coverage, differences in the method adopted, change
94
in the method of reporting etc. Like, if we compare the profit of a concern in March 2000, and then later on
after a few years noticed the profit, there may be a change in the method of charging depreciation, treatment
of revenue items as capital items etc. Such changes would make the comparison meaningless.
Activity B
Take a time series data for 10 years pertaining to production or employee turnover in any organization
familiar to you and point out the preliminary adjustments required.
4.8
Methods of Finding Trends
There are various methods for finding trends which are given below:Measurement of Secular Trend/ Methods of Finding Trend:
Various methods that can be used for determining trend are:
1) Freehand or Graphic Method
2) Statistical (Semi Average Method)
3) Algebraic (Least Square) Method
4) Moving Average Method
1) Free hand curve method: - Plot points on the graph which represents the trend in the time series. Draw
smooth and free hand line among them.
Merits
This method is very simple to apply as it does not involve any calculations.
It is flexible as no rigid boundaries are framed for its application. This method may be applied even
if some data is missing as approximate curve may be obtained.
It saves the time of the researcher as trend may be analyzed without making any cumbersome
calculations.
Demerits
The usage of this method leads to lack of exactness in the estimated trend.
This method is highly subjective as the trend line obtained depends on the personal judgment of the
researcher.
Illustration: From the figures given below, fit a trend line by free hand curve method:
Year
Production
Year
(000 tones)
Production
(000 tones)
1997
2002
19
1998
15
2003
14
1999
10
2004
23
2000
18
2005
18
2001
12
2006
25
95
Solution:
Trend Line
2) Method of Semi Averages: - Under this method the time series data is divided in two equal parts.
The values of both parts are added and their mean is worked out. Plot each on the middle year of each
part and obtain trend line and extend it both sides. If number of years are odd then middle year is left and
remaining values are divided in 2 parts.
Merits
This method is simple to apply and involves less calculations as compared to other methods.
It is objective method as same results are obtained even if it is applied by different researchers.
Demerits
Limitations of arithmetic average apply.
If odd number of years is there, then results are not accurate as the middle year is ignored.
It assumes straight line relationship between plotted points which may be unrealistic in some cases.
Illustration: Fit a trend by the method of semi-averages to the following data:
Year
1993
1994
1995
1996
1997
1998
Sales (Rs. in lakh)

15
22
20
11
27
31
Year
1999
2000
2001
2002
2003
Sales (Rs. in lakh)

28
20
34
30
23
Solution: In the above problem, 11 years are given as such we will leave the 6th year and the
remaining 5-5 years will be taken for calculating the averages.
Year
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
Calculation of Semi-Average
Sales (Rs. in
Total
Semi-Average
lakh)
15
22
20
95
95/5 = 19
11
27
31
28
20
34
135
135/5 = 27
30
23
96
Median Year
1995
2001
34
30
Sales (in lakh
Rupees)
26
22
18
---- Trend Line
14
___ Original Data
10
False Base Line
1992 93
94
95
96
97
98
99
2000
01
02
03
Years
3) Method of least Squares or fitting a straight line trend: - The line obtained by the method of least
squares is called Line of best fit. The sum of the squares of the deviations of the actual and computed
values is least from this line.
Conditions:
Apply Formula [2 eq.], get a & b and put in Yc = a + bx and get trend values.
After eliminating trend, we are left with Cyclical & irregular variations.
Monthly increase = b/12
Illustration: The following table represents profits of a multinational company. Fit a straight line trend and
estimate the figure of profit of the year 2004:
Years
1997
1998
1999
2000
2001
2002
2003
Profit (000 Rs.)
300
700
600
800
900
700
1000
Solution:
97
XY
X2
YC
80
84
95.23
90
90
86
104.65
92
184
88
104.54
83
249
90
92.22
94
376
16
92
102.17
99
495
25
94
105.31
92
552
36
96
95.83
Y = 630
X = 21
XY = 1946
X2 = 91
1980
1981
1982
1983
1984
1985
1986
N=7
Apply formula, get a & b and put in Yc = a + bx and get trend values.
After eliminating trend, we are left with seasonal cyclical & irregular variations.
First calculate value of b
b = Nxy xy
Nx2 (x)2
=
7 x 1946 -21 x 630

7 x 91 (21) 2
13622 -13230
637 441
392
196
b = 2
Substituting the values in (1) above
a=
630 -2 x 21
7
84
Therefore, the trend equation is

Yc = 84 + 2x
Therefore the trend value for the year 1980; substituting x = 0 in example.
=
??
Year
84
98
????
x 100
Similarly trend values can be calculate for other years; and the results are shown in the column under Yc.
Note: If annual change in trend values in given by b then monthly change by
or quarterly change
in given by
Similarly, if the monthly change in trend value in b then the yearly change is given by 12b And quarterly
change is given by 4b.
Eliminating the effect of trend, the detrended values (series) are calculated using.
x 100 for each
year. The results are shown in last column.

Here we write X = 1, corresponding to first year, 1997, X = 2 for the second year 1998 and so on. The
values of profit are denoted by Y.
Merits and demerits of method of least squares:
Merits: 1) There in no possibility for subjective ness as all calculations are formula based and personal
biasness of the researcher cannot affect the results.
2) The various trend values can be obtained for all time periods covered under the study.
3) Line of best fit is obtained. This line gives more appropriate estimations about the future trend
values.
Demerits: 1) There may be different approaches to this method. Thus it becomes difficult to select.
2) This method involves heavy calculations and so it is quite time consuming.
3) The predictions about future economic values are based only on trend.
4) This method is not flexible as the necessary conditions are pre defined with no
scope for any alterations..
4) Method of Moving Averages: Under this method, the average value for a particular number of years (3, 4, 5 or more) is calculated and it
is considered to be the trend value for the unit of time falling at the middle of the period covered.
This is the most common method used to measure secular trend. While using this method it is important to
select period for moving averages; For Example 3 yearly moving averages, 5 yearly moving averages, 8
yearly moving averages etc. This period can be even or odd. It is calculated as follows:
Firstly we decide average period. Example: 3 years moving average
If period is in odd like 3 years then take the average of the first three values and place this against the middle
year, i.e. 2nd year.
After this, find the average of values leaving the values of the first year and including the value of the fourth
year and place it against the 3rd year. Continue this process till you reach the last value of the series.
These moving average figures are also called trend value.
When the period of moving average is even, i.e. 4, 6, 8 years, the following steps are adopted:
If 4 yearly period, add the first four values and place this against in the middle of the second & third year.
99
After this, find the sum of values leaving the value of the first year and including the value of the fifth year and
place it against in the middle of the third and fourth, complete this process till the end.
Now add the first and the second four yearly totals place it against the 3rd year. Similarly continue this
process till end.
Divide the sums derived thus by twice the period of moving average (8 in this case) respectively to get
moving averages.
Illustration
Fit a trend line from the following data by 3 yearly moving average methods.
Year
2001
Production
9
(000 quintals)
2002
11
2003
10
2004
12
2005
14
2006
8
2007
9
2008
10
2009
9
Solution:
Trend Values by 3 yearly moving averages
Year
Production in
3 Yearly moving
totals
3 yearly moving
averages
2001
2002
2003
2004
2005
2006
2007
2008
2009
9
11
10
12
14
8
9
10
9
30
33
36
31
27
28
-
10
11
12
11.33
10.33
9
9.33
-
Here, we have to calculate three yearly moving averages. For this we first calculate the sum of first three
values 9+11+10=30 and place it against the middle year i.e., the second year 2002. Then leaving out the
first value add the values 11+10+12=33 and place it against the third year i.e., 2003. Thus every time we
leave the first of the three and add the next one. These are called moving totals.
These moving totals are divided by 3 to get the moving averages. Moving average represent the trend
values for the specified years.
Illustration
Find 4 yearly moving averages of the sales figure given below:
Year
Sales
(in 000 Rs.)
2003
90
2004
106
2005
104
2006
100
100
2007
105
2008
102
2009
103
2000
102
2011
110
Solution:
Calculation of 4 yearly moving averages
Year
Sales
4 yearly Totales
i)
(ii) (in 000 Rs.)
(iii)
2003
2004
2005
2006
2007
2008
2009
2010
2011
98
106
104
100
105
102
103
102
110
408
415
411
410
412
417
-
4 yearly centered
total
(iv)
823
826
8221
822
829
-
4 yearly centered
average
( iv/8)
102.8
103.2
102.6
102.7
103.6
-
Here, we have to calculate four yearly moving averages. For this we find 4 yearly moving totals and place
them at the center. These four yearly moving totals are centered by taking two yearly moving totals i.e., the
four yearly totals are added two at a time. First we add 408 + 415 = 823 and place it between 408 and 415
i.e., against the year 2005. Next add 415+411=826 and place it again 2006 and so on. To get the average
we divide these centered totals by 8. These are called 4 yearly centered averages and represent the trend
Value.
Merits and demerits of Moving Average:
Merits:
The method is simple to understand and easy to adopt, as compared with the least square method,
as it does require simple calculation..
The moving average method is quite flexible in the sense that a few more observations may be
added to a given data without affecting the trend values already obtained.
The moving average has the advantages that it follows the several movements of the data and that its
shape is determined by the data rather than the statisticians choice of a mathematical function.
The moving average is particularly effective if the trend of a series is very irregular.
Demerits:
Since the moving average is not represented by a mathematical function, this method cannot be
used in forecasting which is one of the main objectives of trend analysis.
Great care has to be exercised in selecting the period of moving average.
They are useful only when the trend is linear or approximately so.
Thus, analysis of time series assumes great importance in business decision making. The increasing relevance
of business forecasting has led to the development of such techniques. In developed countries, there are
many forecasting agencies which provide professional assistance on these techniques. But in case ofdeveloping
countries like India, there are hardly any forecasting agencies emerging. Although, there is a need of such
agencies as business opportunities are growing progressively in India.
101
Methods for calculation of seasonal Variations

Most of the phenomena in economics and business show seasonal patterns. When data are expressed
annually there is no seasonal variation. However, monthly data frequently exhibit strong seasonal movements
and considerable interest attaches to devising a pattern of average seasonal variation. There are many
techniques available for computing an index of seasonal variation, many of the simpler methods were devised
prior to the development of electronic computers and were designed to sacrifice precision to the development
of electronic computers and were designed to sacrifice precision for ease of computation. Any acceptable
modern method for computing such an index probably will be programmed for a computer solution. The
method should be designed to meet the following criteria.
1. It should measure only the seasonal forces in the data. It should not be influenced by the forces of
the forces of trend or cycle that may be present.
2. It should modify the erratic fluctuation in the data with an acceptable system of averaging.
3. It should recognize slowly changing seasonal patterns that may be present and modify the index to
keep up with those changes.
The following are some of the methods more popularly used for measuring seasonal variations:
1. Method of Simple Average (Weekly, Monthly or Quarterly)
2. Ratio- to- Trend Method
3. Ratio- to Moving Average Method
4. Link Relative Method
1
Method of Simple Average:- Method of simple average is the simplest method of obtaining a
seasonal index. Here are the following steps which are necessary for calculating the index.
1. Arrange the unadjusted data by years and months.
2. Find the totals of January , February, etc.
3. Divide each total by the number of years for which data are given. For example:- if we are
given monthly data for five years then, we shall first obtain total for each month for five
years and divide each total by 5 to obtain an average.
4. Taking the average of monthly averages as 100, compute the percentages of various monthly
1. average as follows:
Monthly average for January
5. Seasonal Index for January=
100
Average of Monthly averages
For example:Consumption ofmonthly electric power in million of kwh for street lighting in a big city during 2003-2007
is given below:
Year
2006
2007
2008
Jan
159
171
183
Feb
140
154
164
March
139
150
160
Apr
125
134
143
May
115
124
134
June
108
118
125
102
July
111
121
129
Aug
122
131
142
Sep
134
144
154
Oct
151
160
172
Nov
162
171
183
Dec
173
183
197
Find out seasonal variation by the method of monthly averages.

Solution:
CONSTRUCTION OF SEASONAL INDICES BY
THE METHOD OF MONTHLY AVERAGES
Consumption of monthly electric power
Months. 2006 2007 2008 Monthly totals Three yearly Seasonal
(1)
Jan
Feb
March
April
May
June
July
Aug
Sep
Oct
Nov
Dec
(2)
159
140
139
125
115
108
111
122
134
151
162
173
(3)
171
154
149
134
124
118
121
131
144
160
171
182
(4)
183
164
160
143
134
125
129
143
154
172
183
197
Total
Average
For 3 years
(5)
513
458
448
402
373
351
361
395
432
483
516
552
5284
440.3
average
(6)
171
152.7
149.3
134
124.3
117
120.3
131.7
144
161
172
184
1761.5
146.8
variation
(7)
116.4
104
101.7
91.8
84.7
79.7
81.9
89.7
98.1
109.6
117.1
125.3
1200
100
The above calculations are explained below:

1.
Column No.5 gives the total for each month for 3 years.
2.
In column No.6 each total of Column No.5 has been divided by 5 to obtain an average for each
month.
3.
The average of monthly averages is obtained by dividing the total of monthly averages by 12.
4.
In column No.7, each monthly average has been expressed as a percentage of the average of
monthly averages. Thus, the percentage for January
= 171/146.8*100 = 116.14
If, instead of monthly data, we are given weekly or quarterly data, we shall compute weekly or quarterly
averages by following the same procedure as explained above.
Merits and Demerits of the Method of Monthly Average:
Merits:1. Simple average method is very suitable when materials are received in uniform lot quantities.
2. Simple average method is very easy to operate.
3. Simple average method reduces clerical work.
Demerits:1. If the quantity in each lot varies widely, the average price will lead to erroneous costs.
103
2. Costs are not fully recovered.

3. Closing stock is not valued at the current assets.
Ratio- to- Trend Method: - This method is used when cyclical variations are absent from the
data, i.e. the time series variable Y consists of trend, seasonal and random components. Using
symbols, we can write
Y = T. S .R
Various steps in the computation of seasonal indices are:
1. Obtain the trend values for each month or quarter, etc, by the method of least squares.
2. Divide the original values by the corresponding trend values. This would eliminate trend values from
the data.
3.To get figures in percentages, the quotients aremultiplied by 100. Thus, we have three equations:
Y/ T 100
.......................................(i)
T. S. R / T 100
.......................................(ii)
S. R 100
.......................................(iii)
Q. Find seasonal variations by the ratio-to-trend method from the data given below:
Year
1st Quarter
2nd Quarter
3rd Quarter
4th Quarter
2006
15
20
18
17
2007
17
26
25
22
2008
20
29
27
24
2009
27
38
34
36
2010
40
46
43
41
Solution: For determining seasonal variation by ratio-to-trend method, first we will determine the trend
for yearly data and then convert it to quarterly data.
CALCULATING TREND BY METHOD OF LEAST SQUARES
Year
Yearly
Yearly
Deviations
Totals
average
from mid-year
XY
X2
Trend
Values
2006
70
17.5
-2
-35
16
2007
90
22.5
-1
-22.5
22
2008
100
25
28
2009
130
32.5
32.5
34
2010
170
42.5
85
40
104
N=5
Y 140
XY 60 X
10
The equation of the straight line trend is Y = a+bY
a = Y/N = 140/5 =56

b=XY/X2=60/10 =12
Quarterly increment= 12/4=3
Merits and Demerits
It is an objective method of measuring seasonal variations. However, it is very complicated and doesnt
work if cyclical variations are present.
3
Ratio- to Moving Average Method: - The ratio to moving average is the most commonly used
method of measuring seasonal variations. This method assumes the presence of all the four components
of a time series. Various steps in the computation of seasonal indices are as follows:
1.
Compute the moving averages with period equal to the period of seasonal variations. This would
eliminate the seasonal components and minimize the effect of random component. The resulting
moving averages would consist of trend, cyclical and random components.
2.
The original values, for each quarter ( or month)are divided by the respective moving average
figures and the ratio is expressed as a percentage, i.e. SR = Y / M. A = TCSR / TCR, where R
and R denote the changed random components.
3.
Finally, the random component R is eliminated by the method of simple averages.
Example 11.3.1 : Calculate the three and five year moving averages of the following
data:
Year
1970
1971
Production 18 19
1972
1973
1974
1975
1976
1977
1978
1979
1980 1981
20
22
20
19
22
24
25
24
25
26
(000 tons)
Year
Prediction
y
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
18
19
20
22
20
19
22
24
25
24
25
26
3 yrs
Moving
Trend
57
61
62
61
61
65
71
73
74
75
-
3 yrs
Moving
Average
19.0
20.3
20.6
20.3
20.3
21.6
23.6
24.3
24.6
25.0
-
105
5 yrs
Moving
Trend
99
100
103
107
110
114
120
125
-
5 yrs
Moving
Average
19.8
20.0
20.6
21.4
22.0
22.8
24.0
25.0
-
Steps of Calculation
1) In Table 1 1.3.1 the figures in col. 3 are obtained as the sum of three consecutive values of col. 2.
Thus the first moving total (M.T.) is 57 = 18 + 19 + 20 and is placed against 1971. The second
moving total 61 = 19 + 20 + 22 is placed against 1972.
2) The three-year moving average(M.A.) in col. 4 is obtained by dividing the corresponding threeyear moving total in col. 3 by 3, the period of the moving average. Thus 57 3= 19, 61 3 =
20.3, etc.
3) The five-year moving totals in col. 5 are obtained as the sum of five consecutive values in col. 2.
Thus the first moving total against the year 1972 is 99 = 18 + 19 + 20 + 22 + 20.
4) The five-year moving average in col. 6 is obtained by dividing the corresponding five-year moving
total in col. 5 by 5,. Thus the moving average for 1975 is 107 5 = 21.4.
Merits and Demerits
This method assumes that all the four components of a time series are present and, therefore, widely used
for measuring seasonal variations. However, the seasonal variations are not completely eliminated if the
cycles of these variations are not of regular nature. Further, some information is always lost at ends of the
time series
4.
Link Relative Method: This method is based on the assumption that the trend is linear and
cyclical variations are of uniform pattern. The link relatives are percentages of the current period
(quarter or month) as compared with the previous period. With the computations of the link relatives
and their average, the effect of cyclical and the random components is minimized. Further, the trend
gets eliminated in the process of adjustment of chain relatives.
The following steps are involved in the computation of seasonal indices by this method:
1. Compute the Link Relative (L.R.) of each period by dividing the figure of that period with the figure
of previous period. For example, Link relative of 3rd quarter=figure of 3 rd quarter /figure of 2nd
quarter 100.
2. Obtain the average of link relatives of a given quarter (or month) of various years. A.M. or M d can
be used for this purpose. Theoretically, the later is preferable because the former gives undue
importance to extreme items.
3. These averages are converted into chained relatives by assuming the chained relative of the first
quarter (or month) equal to 100. The chained relative (C.R.)for the current period (quarter or
month)= C.R. of the previous period L.R. of the current period / 100.
4. Compute the C.R. of the first quarter (or month) on the basis of the last quarter (or month). This is
given by C .R. of the last quarter (month) average L.R. of the first quarter (month) / 100
This value, in general, is different from 100 due to long term trend in the data. The chained relatives,
obtained above, are to be adjusted for the effect of this trend. The adjustment is to be done follows:
a)
Calculate
Chain Relative (C.R.) of 1st month on the basis of last month
CR of last month Average of link Re lative of 1st month

100
106
b)
Find this is k
Difference between the two chain relatives
= 100- k
[this can be +ve or -ve]
c) Make adjustment
Correction factor for each quarter =
100 k
CF
4
Corrected Values:
Corrected C.R. of 1st quarter = 100
Corrected C.R. of 2nd quarter = CR of 2nd quarter 1X CF
Corrected C.R. of 3rd quarter = CR of 3rd quarter 2X CF
Corrected C.R. of 4th quarter = CR of 4th quarter 3X CF
5.
Express the adjusted chained relatives as a percentage of their average to obtain seasonalindices.
6.
Make sure that the sum of these indices is 400 for quarterly data and 1200 for monthly data.
Merits and Demerits

This method is less complicated than the ratio to moving average and the ratio to trend methods. However,
this method is based upon the assumption of a linear trend which may not always hold true.
4.9
Summary
Business forecasting refers to the estimation of proposed business situations. It helps in proper planning,
optimum utilization of resources, increase in productivity and systematic working in the business. There are
various methods which may be applied for business forecasting viz. time series analysis, opinion poll method,
survey method, regression analysis, econometric method etc.
Time series is the most prominently used method which helps in identifying the type of fluctuations in the
business activities i.e. Secular, Cyclical, Seasonal or Irregular. There are various methods of working out
these fluctuations. These are free hand method, statistical, algebraic and moving average method. The most
widely applied method is moving average as it rules out the effect of extreme values in the data. These
techniques are of immense use in business as it helps to reduce the risk and uncertainty associated with
future.
4.10 Key Words
Business: It refers to an organization or enterprising entity engaged in commercial, industrial or

professional activities. A business can be a for-profit entity, such as a publicly-traded corporation,
or a non-profit organization engaged in business activities.
Business Forecasting: It is concerned with trying to reduce the uncertainty relating to future.
Additional information is generated about the future that help the managers to assess the future
consequences of existing decisions and to evaluate the consequences of alternatives.
107
Time Series: A time series is a sequence of data points, measured typically at successive time
instants spaced at uniform time intervals.
Variables: Those entities whose value are not stable but keep on changing are called variables.
4.11 Self Assessment test

1.
What do you understand by Business Forecasting? Explain the objectives underlying the
business forecasting?
2.
Discuss the methods of Business Forecasting?
3.
Define Time Series? State the main components of Time Series?
4.
Explain the preliminary adjustments required before analyzing time series data?
5.
State the equations to find out the values of a and b constants in the trend equation:
Yc = a + bX
6.
State the difference between seasonal variations and cyclical fluctuations?
7.
Apply the semi average method for measuring trend to the following data:
1970 1971
Years
Production 102
8.
105
1972 1973 1974
1975 1976 1977 1978 1979
1780 1981
114
116
112
110
108
112
120
114
108
119
Calculate the trend value by 3 yearly moving averages from the following data:
Year
2001
2002
2003
2004
2005
2006
2007
2008
2009
Sales
16
14
20
18
22
17
19
21
20
4.12 References
Mik Wisniewski (2002). Quantitative Methods for Decision Makers, Macmillian India Ltd.
Schaums Outline of Business Statistics, Leonard J. Kazmier, Fourth Edition (2009), Tata
McGraw Hill
Basic Business Statistics, Mark L. Berenson (6th edition) Pearson New Delhi.
David Freedman, Robert Pisani, Roger Purves, Statistics (Fourth Edition).
108
Unit - 5 : Index Numbers

Structure Unit:
5.0
Objectives
5.1
Introduction
5.2
Uses of Index Numbers
5.3
Types of Index Numbers
5.4
Precautions in Construction of Index Numbers
5.5
Notations and Methods of Constructing Index Numbers
5.6
Test of Consistency
5.7
Fixed and Chain Base Index Numbers
5.8
Base Shifting, Splicing and Deflating the Index Numbers
5.9
Limitations of Index Numbers
5.10 Summary
5.11 Key Words
5.13 References
5.0
Objectives
After completing this unit, you will be able to:

Understand that index numbers describe how much economic variables have changed over time;
Point out various uses of index numbers;
Know and avoid problems in constructing index numbers;
Become familiar with the three principal types of index: price index, quantity index, and value index;
Learn how to calculate various kinds of index numbers;
Describe various limitations of index numbers.
5.1 Introduction
In our day to day life, things keep changing. The prices of various commodities vary at some rate over a
period of time. In order to be update on such price changes, we need some methods to predict these. In
business area for the budget purpose a manager may be interested to know how the raw material prices
have increased over last one year. May be some changes in price; indicate the trend of increase on the
regular basis, so that future planning could be more accurate. For such variations, we may analyze the
degree of change in the form of Index Numbers.
Index numbers are indicators which reflect the relative changes in the level of a certain phenomenon in any
given period (or over a specified period of time) called the current period with respect to its values in some
fixed period, called the base period selected for comparison.
Index numbers are often known as the barometers of economic activity as they help to get an idea of
the present day situation with regard to changes in production, consumption, exports and imports, national
income, business level, cost of living, the price of a particular commodity or a group of commodities,
Industrial or agricultural production, stocks and shares, sales and profits of a business house, volume of
trade, factory production, wage structure of workers in various sectors, bank deposits, foreign exchange
reserves, and so on.
Some of the important definitions of index numbers are given as under:
According to Wessell, Willet and Simone, An index number is a special type of average that provides
a measurement of relative changes from time to time or from place to place.
109
According to Edgeworth, Index number shows by its variations the changes in a magnitude which is not
susceptible either of accurate measurement in itself or of direct valuation in practice.
According to Murray R. Spiegel, An Index Number is a statistical measure designed to show changes
in variable or a group of related variables with respect to time, geographic location or other characteristics.
According to Croxton and Lowden, Index Numbers are devices for measuring differences in the magnitude
of a group of related variables.
According to Karmal and Pollasek, An index number is a device for comparing the general level of
magnitude of a group of distinct, but related, variables in two or more situations.
According to Patterson, In its simplest form, an index number is the ratio of two index numbers expressed
as a percent. An index number is a statistical measure - a measure designed to show changes in one variable
or in group of related variable over time or with respect of geographic location or other characteristics.
According to Dr. A. L. Bowley, A series of index numbers is a series which reflects in its trend and
fluctuations, the movements of some quantity to which it is related.
According to John I. Griffin, An Index Number is a quantity which by reference to a base period
shows by its variations the changes in the magnitude over a period of time. In general, Index Numbers are
used to measure changes over time in magnitudes which are not capable of direct measurement.
Thus, it is apparent from above definitions that an index number is a statistical device which measures the
extent to which a group of related variable changes over a period of time. Index number in fact relates a
variable or group of variables in a given period to the same group of variables in some other period.
Some of the important characteristics of index numbers include the following:
Index numbers are the specialized averages.
Index numbers record the net changes in a group of related variables over a period of time.
Index numbers measure changes not capable of direct measurement.
Index numbers are for comparison.
Index numbers are expressed in percentages.
Activity A:
1. An Index number is a special type of average. Discuss.
2. According to you which definition of index number is more appropriate and why?
5.2
Uses of Index Numbers
The first index number was constructed by an Italian, Mr. Carli, in 1764 to compare the changes in price for
the year 1750 (current year) with the price level in 1500 (base year) in order to study the effect of price level
in Italy. Though originally designed to study the general level of prices or accordingly purchasing power of
money, today index numbers are an important tool of economic analysis and they reveal the pulse of the
economy.
Use of Index numbers is the most powerful tool in the hands of management, government officials and
individuals to analyse the business and economic situations of a country. Some of the important uses or
significance of index numbers to its users are listed as under:
1. Index Numbers Help in Formulating Policies and Decision Making: Formulation of good policies
for the future depends upon past trends. Behaviour of the index numbers is studied carefully before making
any policies. Index numbers of the data relating to prices, production, profits, imports and exports, personnel
and financial matters are indispensable for any organisation in efficient planning and formulation of executive
decisions.
For example the cost of living index numbers help the employers in deciding about the increase in dearness
110
allowance of their employees or adjusting their salaries and wages in accordance with changes in their cost
of living.
2. Reveal Trends and Tendencies: The index numbers study the relative changes in the level of a
phenomena. So, they would disclose the general trend for a variable or group of variables in time series
data. For example by examining the index numbers of production (industrial and agricultural), volume of
trade, imports and exports etc. for the last few years, we can draw useful conclusions about the trend of
production and business activity.
3. Index Numbers Measure the Purchasing Power of Money: Index numbers are helpful in finding out
the intrinsic value of money as contrasted with its nominal worth. The cost of living index numbers determine
whether the real wages are rising or falling, money wages remaining unchanged.
For example, suppose that the cost of living index for any year, say, 2010 for a particular class of people
with 2001 as base year is 200. If a person belonging to that class gets Rs. 300 in 2001, then in order to
maintain the same standard of living as in 2001 (other factors remaining constant) his salary in 2010 should
be
= Rs.600.
4. Aid in Deflation: Index numbers are very useful for deflating (or adjusting) the original data. In obtaining
real income from inflated income, real wages from nominal wages, and real sales from nominal sales and so
on, the index numbers are immensely useful.
5. They are Economic Barometers. As described above various index numbers are computed for different
purposes, say employment, trade, transport, agriculture, industry, etc., and these are of immense value in
dealing with different economic issues. Like barometers which are used in Physics and Chemistry to measure
atmospheric pressure, index numbers are rightly termed as economic barometers or barometers of
economic activity which measure the pressure of economic and business behaviour.
In the words of G. Simpson and F. Kafka, Index numbers are today one of the most widely used
statistical devices. They are used to take the pulse of the economy and they have come to be used as
indicators of inflationary or deflationary tendencies.
6. Other Uses: Index numbers are thus, become indispensable tool in business planning and formulation of
executive decisions of any concern and theses are today one of the most widely used statistical devices.
Fairly good appraisal of trade and business activities of country can be done if careful study is carried out by
constructing index. In brief the uses of index numbers are shown below:
They measure the relative change
They are of better comparison.
They are wage adjuster.
They compare standard of living.
They are special type of averages.
Activity B:
1. Index numbers are the devices for measuring differences in the magnitude of a group of
related variables. Discuss this statement and point out the important uses of index numbers.
5.3
Types of Index Numbers
Index numbers may be divided into three categories:

(a) Price Index
(b) Quantity Index, and
(c) Value Index
(a) Price Index: It is the most commonly used index. It compares the prices of various commodities from
111
one period to another. We may have steel price index, sugar price index, vegetables price index etc. For
these purpose, a well known price index is called consumer price index (CPI), it is tabulated at regional or
National level so as to establish the price levels of various consumer goods and services. This is an effective
measure and useful index of cost of living. Index numbers are further divided into the following two categories:
(i) Wholesale Price Index Numbers: It reflect the changes in general price level of the country
such as wholesale price index number prepared by government.
(ii) Retail Price Index Numbers: These index numbers show general changes in retail prices of various
commodities such as consumption goods, stock and shares, bank deposits, consumer price index etc.
(b) Quantity Index: Quantity index numbers study the changes in the volume of goods produced, distributed
or consumed, like the indices of agricultural production, industrial production, imports and exports, etc.
These types of indices are useful for measuring the changes in level of physical output in an economy during
some period compared with other period.
(c) Value Index: The last type of Index, the value index, measures changes in total monetary worth, during
some period compared with other period. The value index numbers are intended to study the changes in the
total value (quantity multiplied by price) of production.
5.4
Precautions in Construction of Index Numbers
Index numbers which are not properly compiled will, not only lead to wrong and fallacious conclusions but
might also prove to be dangerous. So, the construction of index number requires a careful study of some
aspects which often called precautions or problems. The given below are some precautions:
1. Purpose of Index Numbers: It is essential to be clear about the purpose for which the index number is
used. Every index number has its own particular uses. For example if it is used for measuring consumers
price, there is no need of including wholesale prices. Similarly, if it is employed for studying cost of living,
there is no need of including the price of capital goods.
2. Selection of Base Period: Base period is the period against which comparisons are made. One has to
be very careful in selecting a base period. If we select inappropriate base, then distortion can be large.
Taking a year of large increase in prices due to abnormal causes may not be appropriate base. Also may be
a large consumption of a commodity at the time of natural calamity may not reflect correct index level at the
base year. Before selecting a base period the following points should be kept in mind:
(a) The base period should not get affected by extra-ordinary events like war, earthquakes, famines,
booms, etc. it should be a normal one
(b) It should be relatively current i.e., it should not be too distant in the past because we are
interested in the changes relating to the present period only.
(c) The base may be fixed, chain or average depending upon the purpose of constructing the index.
3. Selection of Commodities or Items: Only those commodities or items which are fairly represented
and uniform quality should be selected for inclusion in the construction of index number. White selecting the
sample the following points should be kept in mind:
The commodities selected should be relevant to the purpose of the index.
Select the adequate number of representative items from each group (Neither too small nor too large)
Classify the whole relevant group of items or commodities into relatively homogeneous sub-groups.
4. Selection of Weights: Weights imply the relative importance of the different variables. Due to
inappropriate importance given to various factors in calculation of index, the calculated value willbe found
distorted and may not be the true representative of the decision variable. So proper weight should be
assigned to different commodities with their relative importance in the group. Assigning weights canbe done
by (a) Implicit method; items are implicitly weighted, assumed to equal importance) and (b) Explicit method
112
(weights are assigned either in quantity terms or in value terms). The choice of method of weighing depends
on the purpose, scope, availability of data. Like in developing a composite index, such as the Consumer
Price Index, we must consider changes in some variables to be more important than changes in others like,
wheat should be given more importance compared to sugar.
5. Data Collection: The basic data used must be reliable, authentic and suitable for the purpose. Sufficient
data is also required to arrive at a worthwhile information deduction. The source of data depends on their
information requirements. In dealing with broad areas of national economy and the general level of business
activity, publications such as the Federal Reserve Bulletin, Moodys, Monthly Labour Review, and the
Consumer Price Index provide a wealth of data. Almost all government agencies distribute data about their
activities, from which index numbers can be computed. Many financial newspapers and magazines provide
information from which index numbers can be computed. When you read these sources, you will find that
many of them use index numbers themselves.
6. Selection of Average: Averages play a vital role in arriving at a single index number summarizing a large
volume of information. Arithmetic mean and geometric mean are used in it construction but theoretically, the
geometric mean is preferred because it is less susceptible to variation; it gives equal weight to equal ratio of
change.
7. Price Collection: After selecting the items, the next problem is to collect their prices. The price of a
commodity varies from place to place and even from shop to shop in the same market. So, it is very difficult
to consider and compile prices from every market, from every shop and for all periods. Therefore, we
should select, a sample market, which are well known for trading in a particular commodity and also collect
the data of price for that commodity from the agencies such as the Chambers of Commerce, News
Correspondents etc. and further we should compare for validity and reliability.
8. Selection of Appropriate Formula: A large number of formulas have been devised for constructing the
index numbers. A decision has, therefore, to be made as to which formula is the most suitable for the
purpose. The choice of the formula depends upon the availability of the data regarding the prices and
quantities of the selected commodities in the base and/or current year.
Activity C:
1. What is an Index number? Why index numbers are called Economic barometers?
2. Analyze the problems in the construction of index numbers and comment.
5.5
Notations and Methods of Constructing Index Numbers
Notations:
Base Year: The year selected for comparison i.e. the year with reference to which comparisons are made.
It is denoted by 0.
Current Year: The year for which comparisons are sought or required.
P0
P1
q0
q1
W
P01
P10
Q 01
q10
is price of a commodity in the base year,

is price of a commodity in the current year,
is quantity of a commodity in the base year,
is quantity of a commodity in the current year,
is weight assigned to a commodity according to its relative importance in group,
is price index number for the current year,
is price index number for the base year,
is quantity index number for the current year, and
is quantity index number for the base year
113
Methods of Constructing Index Numbers

(I) Price Index Numbers
Methods of constructing index numbers can broadly be divided into two classes namely:
(A) Un-weighted Index Numbers, and
(B) Weighted Index Numbers.
In case of un-weighted indices, weights are not assigned, whereas in the weighted indices weights are
assigned to the various items. Each of these types may be further classified under two heads:
(i) Aggregate of Prices Method, and
(ii) Average of Price Relatives Method.
The following chart illustrates the various methods:
Index Numbers
Un-weighted
Simple Aggregate
of prices
Weighted
Simple Average
of price relatives
Weighted Aggregate
of prices
Weighted Average
of price relatives
(A) Un-weighted Index Numbers:

(i) Simple (Un-weighted) Aggregate Method: This is the simplest methods of constructing index numbers
and consists in expressing the total price, i.e., aggregate of prices (of all the selected commodities) in the
current year as a percentage of the aggregate of prices in the base year. Thus, the price index for the current
year w.r.t. the base year is given by:
Where
is the aggregate of prices in the current year and
is the aggregate of prices in the base year..
This method suffers from a drawback that equal weight is given to all the items irrespective of their relative
importance.
Illustration: 1
From the following data calculate Index Number by simple aggregate method.
Commodity
:
A
B
C
D
Price in 1990 (Rs.) :
162
256
257
132
Price in 1991 (Rs.) :
171
164
189
145
Solution:
A
B
C
D
Total
Computation of Price Index Number

Price (in Rupees)
1990 (P0)
1991(p1)
162
171
256
164
257
189
132
145
p0=807
p1=669
The price index number using simple aggregate method is given by:
=
= 82.90
114
Illustration: 2
Following prices are indicated for 2005 (base year) and for 2010 (the current year). Calculate the unweighted aggregates price index for the data
Prices
Variables
Tomatoes (per kg.)
Egg (per dozen)
Petrol (Per Liter)
Juices (Per Liter)
2005
2010
Rs. 15.00
Rs. 20.00
Rs. 22.50
Rs. 61.00
Rs. 19.00
Rs. 24.00
Rs. 30.70
Rs. 69.00
Solution:
For computing price index (un-weighted) we have
= Rs. 19+24+30.70+69 = Rs. 142.70
= Rs. 15+20+22.50+61=Rs. 118.50
Un-weighted aggregates price index
(ii) Simple Average of Price Relatives: In order to calculate, we can compare the ratio of current prices
to the base prices and then the index is calculated by multiplying the rate by 100. We, then, take the average
of all the ratios summed up together. The general relationship now undergoes the change as follows:
Un-weighted average of relative price index
The simple average of price relatives method is superior to the simple aggregate of prices method in two
respects:
(i) Since we are comparing price per liter with price per liter, and price per kilogram with price per
kilogram the concealed weight due to use of different units is completely removed.
(ii) The index is not influenced by extreme items as, equal importance is given to all items.
The greatest drawback of unweighted indices is that equal importance or weight is given to all items included
in the index number which is not proper. As such, unweighted indices are of little use in practice.
Illustration: 3
From the following data, construct index number by simple average of price relatives using arithmetic mean
Commodity
Wheat
Rice
Milk
Eggs
Sugar
1998 Price (Rs.)

800/quintal
15/Kg.
12/Liter
10/Dozen
14/Kg.
Solution:
115
2001 (Price (Rs.)

1000/quintal
19/Kg.
15/Liter
12/Dozen
18/Kg.
Commodity
Wheat
Rice
Milk
Eggs
Sugar
P0
800
15
12
10
14
P1
1000
19
15
12
18
P (Price relative)
1000 x 100/800 =
125
19 x 100/15 = 126.67
15 x 100/12 =
125
12 x 100/10 =
120
18 x 100/14 = 128.57
625.24
Using simple Arithmetic Mean:

Average of relative price index
=
=
(B) Weighted Index Numbers:

The purpose of weighting is to make the index numbers more representative and to give more importance to
them. Weighted index numbers are of two types. They are:
(i) Weighted Aggregate Index Numbers. According to this method, prices themselves are weighted by
quantities; i.e., p x q. Thus physical quantities are used as weights. Here are various methods of assigning
weights, and thus various formulas have been formed for the construction of index numbers. Some of the
important formulae are given below:
1. Laspeyres Method
2. Paasches Method
3. Dorbish and Bowleys Method
4. Marshall-Edgeworth Method
5. Kellys Method, and
6. Fishers Ideal Method
1. Laspeyres Method. In this method, base year quantities are taken as weights. The formula for
constructing the index is:
P01 =
Where P 1
= Price in the current year
P0
= Price in the base year
q0
= Quantity in the base year
According to this method, the index number for each year is obtained in three steps:
(a) The price of each commodity in each year is multiplied by the base year quantity of that commodity.
For the base year, each product is symbolised by P0q0, and for the current year by P1q0.
(b) The products for each year are totaled and P1q0 and P0q0 are obtained.
(c) P1q0 is divided by P0q0 and the quotient is multiplied by 100 to obtain the index.
Laspeyres index is very widely used in practice. It tells us about the change in the aggregate value of base
period list of goods when valued at given period price. However, this index has one drawback. It does not
take into consideration the changes in the consumption pattern that take place with the passage of time.
116
2. Paasches Method. In this method, the current year quantities are taken as weights: symbolically,
P01(Pa)
According to this method, the index number for each year is obtained in following steps:
1. Multiply current year prices of various commodities with current year weights and obtain P1q1.
2. Multiply the base year prices of various commodities with the current year weights and obtain
P0 q 1 .
3. P1q1 is divided by P0q1 and the quotient is multiplied by 100 to obtain the index
Although this method takes into consideration the changes in the consumption pattern, the need for collecting
data regarding quantities for each year or each period makes the method very expensive. Hence, where the
number of commodities is large, Paasches method is not used in practice.
3. Bowley Dorbish Method. This is an index number got by the arithmetic mean of Laspeyres and
Paasches methods; symbolically (This method takes into account both the current and the base periods).
Symbolically
L = Laspeyres method
P = Paasches method
4. Marshall-Edgeworth Method. In this method, the totals of base year and current year quantities are
taken as weights. The formula for constructing the index is:
5. Kellys Price Index or Fixed Weights Index. This formula, named after Truman L. Kelly, requires the
weights to be fixed for all periods and is also sometimes known as aggregative index with fixed weights
and is given by the formula:
Where the weights are the quantities (q) which may refer to some period (not necessarily the base year or
the current year) and are kept constant for all periods. The average (A.M. or G.M.) of the quantities
consumed of two, three or more years may be used as weights.
Kellys fixed base index has a distinct advantage over Laspeyres index because unlike Laspeyres index
the change in the base year does not necessitate a corresponding change in the weights which can be kept
constant until new data become available to revise the index. As such, currently this index is finding great
favour and becoming quite popular.
6. Fishers Ideal Index. This method is the geometric mean of Laspeyres and Paasches indices. The
formula for constructing the index is :
117
Because of the following reasons, Fishers formula is known as ideal:

It takes into account prices and quantities of both current year as well as base year.
It uses geometric mean which, theoretically, is the best average for constructing index numbers.
It satisfies both time reversal test and the factor reversal test.
It is free from bias. The weight biases embodied in Laspeyres and Paasches methods are crossed
geometrically and thus eliminated completely.
Illustration: 4
By using Laspeyres method, calculate the weighted price index for the year 2010 when the given data
indicates the prices and consumption levels of various commodities.
Commodities
Base Price
(2007)
Rs.5.10
Rs.14.00
Rs.21.00
Rs.17.50
Potatos (per kg)

Milk (per litre)
Eggs (per doz)
Bread (per loaf)
Current Price
(2010)
Rs.4.50
Rs.17.00
Rs.24.00
Rs.19.00
Average quantity consumed

(2007)
4000 Kgs
800 litres
2000 dozens
350 loaves
Solution:
For working out Laspeyres Price Index, we prepare the table as follows:
Commodities
Potatos (per
kg)
Milk (per litre)
Eggs (per doz)
Bread (per
loaf)
Price in
2007
(P0)
Rs.5.10
Price in
2010
(P1)
Rs.4.50
Quantity in
2007
(Q0)
4000 Kgs
P0Q0
20,400
P1Q0
18,000
Rs.14.00
Rs.21.00
Rs.17.50
Rs.17.00
Rs.24.00
Rs.19.00
800 litres
2000 dozens
350 loaves
11,200
42,000
6,125
13,600
48,000
6,650
From the above calculations
P0Q 0 = Rs. 79,725

P1Q 0 = Rs. 86,250
Laspeyres Price Index = =
x 100
=
= 108
Illustration: 5
From the following data, construct the Laspeyres, Paasches and Fishers indices of prices:
Item
A
B
C
D
Base Year
P0
4
8
2
10
Current Year
q0
20
4
10
5
118
P1
10
16
4
20
q1
15
5
12
6
Solution:
Calculation of Price Index Numbers
Base Year Current Year
P0
q0
P1
q1
P 0 q0
P1q0
P0q1
P 1 q1
4
20
10
15
80
200
60
150
A
8
4
16
5
32
64
40
80
B
2
10
4
12
20
40
24
48
C
10
5
20
6
50
100
60
120
D
Total
P0q0=182 P1q0=404 P0q1=184 P1q1=398
Item
Laspeyres Price Index (L)
Paasches Method (P)
Fishers Ideal Index
=
=
Illustration: 6
From the following data, calculate the price index numbers for 2008 with 2000 as base by:
(a)
Laspeyres method
(b)
Paasches method
(c)
Bowley method
(d)
Marshall-Edgeworth method
(e)
Fishers Ideal method
2000
Price (Rs.)
Quantity (unit)
70
28
175
35
140
52.5
70
70.0
Item
Maize
Millet
Sugar
Coconut
2008
Price (Rs.)
Quantity (unit)
140
21.0
210
17.5
175
52.5
70
87.5
Solution:
Calculation of Price Index Numbers
Item
P0
70
Maize
Millet 175
Sugar 140
Coconut 70
q0
28
35
52.5
70
(a)
Laspeyres method
(b)
Paasches method
P1
140
210
175
70
q1
21.0
17.5
52.5
87.5
P 1 q0
3920
7350
9187.5
4900
25357.5
119
P 0 q0
1960
6125
7350
4900
20335
P1q1
2940
3675
9187.5
6125
21927.5
P0q1
1470
3062.5
7350
6125
18007.5
(c)
Bowley method
(d)
Marshall-Edgeworth method
(e)
Fishers Ideal method
Activity D:
From the data given below, construct index number of prices for 2006 with 2000 as base,
using
(i) Laspeyres method,
(ii) Paasches method,
(iii) Bowley-Drobisch method,
(iv) Marshall - Edgeworth method, and
(v) Fishers ideal formula.
Commodity
A
B
C
D
2000
Price per unit Expenditure
in rupees
2
10
3
12
1
8
4
20
2006
Price per unit Expenditure
in rupees
4
16
6
18
2
14
8
32
(ii) Weighted Average of Price Relatives: This method is similar to the simple average of price relatives
method with the fundamental difference that explicit weights are assigned to each commodity included in the
index. Since price relatives are in percentages, the weights used are value weights. The following steps are
taken in the construction of weighted average of price relatives index:
(i) Calculate the price relatives,
, for each commodity
(ii) Determine the value weight of each commodity in the group by multiplying its price in base year
by its quantity in the base year, i.e., calculate P0q0 for each commodity. If, however, current year
quantities are given, then the weights shall be represented by P1q1.
(iii) Multiply the price relative of each commodity by its value weight as calculated in above (ii).
(iv) Total the products obtained under (iii) above.
(v) Divide the total (iv) above by the total of the value weights. Symbolically index number obtained
by the method of weighted average of price relatives is:
120
Illustration: 7
Based on the data given in illustration 4 calculate the weighted average of relatives index.
Solution:
Calculation of Weighted Average of Relatives Index
Commodities
(P0)
(P1)
Potatos
Milk
Eggs
Bread
Rs.5.10
Rs.14.00
Rs.21.00
Rs.17.50
(Q0)
Rs.4.50 4000
Rs.19.00 800
Rs.24.00 2000
Rs.19.00 350
P0Q0
88.23
135.71
114.28
108.57
20,400
11,200
42,000
6,125
17,99,892
15,19,952
47,99,760
6,64,991
From the above calculations, we obtain the values
P0Q0 = Rs.79,725
and
Weighted Average of the relatives index
(II) Quantity Index Numbers:

A quantity index number is a statistical device which measures changes in quantities in current year as
compared to base year. Quantity index numbers reflect the relative changes in the quantity or volume of
goods produced, consumed, marketed or distributed in any given year w.r.t. to some base year.
The formulae for calculating the quantity index numbers can be directly written from price index numbers
simply by interchanging the role of price and quantity.
Thus quantity index by different methods is:
(a)
Laspeyres method
(b)
Paasches method
(c)
Fishers Ideal method =
Illustration: 8. Compute quantity index for the year 1992 with base 1990=100, for the following data,
using (i) Laspeyres method (ii) Paasches method. (iii) Fishers ideal formula.
121
Price
Item
Quantities
1990
5.00
7.75
9.63
12.50
A
B
C
D
1992
6.50
8.80
7.75
12.75
1990
5
6
4
9
1992
7
10
6
9
Solution:
Computation of Quantity Index
Commodity
A
B
C
D
P0
5.00
7.75
9.63
12.50
q0
5
6
4
9
P1
6.50
8.80
7.75
12.75
(i)
Laspeyres quantity index or Q01
(ii)
Paasches quantity index or Q01
(iii)
Fishers quantity index or Q01
q1
7
10
6
9
q0 P 0
25.00
46.50
38.52
112.50
q0P0 =
222.52
q0P1
32.50
52.80
31.00
114.75
q0P1 =
231.05
q1P0
35.00
77.50
57.78
112.50
q1P0 =
282.78
q1 P 1
45.50
88.00
46.50
114.75
q1P1 =
294.75
= 1.273 x 100
= 127.3
(III) Value Index Numbers:

These index measures the changes in the total value of the variable. Since value is a combination ofprice and
quantity, it can be called a composite index. The only negative issue is that composite value index does not
distinguish the variations in individual values of price or quantity separately. Value index numbersare obtained
on expressing the total value (or expenditure) in any given year as a percentage of the same in the base year.
Symbolically, we write
The value index number based on the information given in illustration 8 can be calculated as under:
Value Index number =
5.6
Test of Consistency
As there are several formulae for constructing index numbers the problem is to select the most appropriate
formula in a given situation. Prof. Irving Fisher has suggested two tests for selecting an appropriate formula.
122
These are:
(A) Time Reversal Test, and
(B) Factor Reversal Test
(A) Time Reversal Test:
This test requires that the formulae for calculating an index number should give consistent results in both the
directions, i.e. forward and backward. Or in other words, the index of period 1 with period 0 base should
be reciprocal of the index of period 0 with period 1 as base i.e. P01 = 1/P10 or p01 x P10 = 1.
This test is satisfied by the Fishers Ideal Index.
We can write
(B) Factor Reversal Test

This test requires that the product of price index and the corresponding quantity index numbers should be
equal to the value index number i.e. P01xQ01=V01.
This test is also satisfied by the Fishers ideal index.
We can write
and
Illustration: 9
Compute Index Number, using Fishers Ideal formula and show that it satisfies time-reversal test and factorreversal test.
A
B
C
D
Quantity
12
15
24
5
Base Year Price

10
7
5
16
123
Quantity
15
20
20
5
Current year Price

12
5
9
14
Solution:
Computation of Index Number
Commodity
A
B
C
D
P01
q0
12
15
24
5
p0
10
7
5
16
q1
15
20
20
5
p1
12
5
9
14
p1 q0
144
75
216
70
p1q0
= 505
p0 q0
p1 q1
p0 q1
120
180
150
105
100
140
120
180
100
80
70
80
p0q0 p1q1 = p0q1
= 425
530
= 470
x 100
x 100
(a) Time-Reversal Test

Time Reversal Test is satisfied when P01 x P10 = 1
P10
P01 x P10
=1
(b) Factor-Reversal Test
Factor-Reversal test is satisfied when
P01
Q01
P01xQ01
P01xQ01
Hence the given data satisfies the time-reversal test and factor-reversal test.
124
Activity E:
1. What are the tests of a good index number? Define Fishers Ideal index number and show
that it satisfies all these tests.
2. State whether the following statements are True or Untrue:
(i) Arithmetic mean is the most appropriate average for constructing the index numbers.
(ii) Paasches Index number is based on base year quantity.
(iii) Fishers Index Number is an Ideal Index Number.
(iv) Fishers Index Numbers is the arithmetic average of Laspeyres and Paasches Index
Numbers.
(v) Time reversal test is satisfied by both formulas Fisher and Kellys.
5.7
Fixed and Chain Base Index Numbers
Fixed Base Index Numbers:

When the comparison of (prices or quantities etc.) various periods are done with reference to a particular or
fixed period, we get an index number series with fixed base.
Chain Base Index Numbers:
The main problem with a fixed base series arises when the current year becomes too far off from the base
year. In such a situation, it may happen that the commodities which used to be very important in the base
year are no longer so in current year. Furthermore, certain new commodities might be in use while some old
commodities are dropped in current year. This problem is often solved by constructing Chain Base Index
Numbers. A chain base index number is an index number with previous year as base.
Differences between Chain Base Method and Fixed Base Method
S.
Chain Base
No.
1. The base year changes.
2. Here the link relative method is used.
3. Introduction and deletion of items are
easy to calculate, without recalculation
of the entire series.
4. The calculations are tedious
5. It is difficult to understand.
The base year does not change.

No such link relative method is used.
Any change in the commodities, will
involve the entire index number to be
recast.
The calculations are simple.
It is simple to understand.
6.
There is no such problem.
7.
8.
9.
It cannot be computed if data for any

one year are missing.
It is suitable for short period only.
Weights can be adjusted as frequently
as possible.
Index number is wrong if an error is
committed in the calculation of any link
index number.
Fixed Base
It is suitable for long periods only.

Weights cannot be adjusted so
frequently.
This is not so, the error is confined to
the index of that year only.
Illustration: 10. From the following data relating to the wholesale prices of wheat for six years, construct
index numbers using (a) 1990 as base, and (b) by chain base method.
125
Year
1990
1991
1992
Price (per quintal)

Rs.
100
120
125
Price (per quintal)

Rs.
130
140
150
Year
1993
1994
1995
Solution:
(a) Computation of Index numbers with 1980 as base:
Year
Price of
wheat
Index Number
1980 = 100
Year
Price of
wheat
1990
100
100
1993
130
1991
120
1994
140
1992
125
1955
150
Year
Price of
wheat
1993
130
Index Number
1980 = 100
(b) Construction of Link Relative Indices

Year
Price of
wheat
Link Relative Index
1990
100
100
1991
120
1994
140
1992
125
1995
150
Link Relative Index
Conversion of Link Relatives into Chain Relatives:

Chain relatives or chain indices can be obtained either directly or by converting link relatives into chain
relatives with the help of the following formula:
Chain relative for current
year =
Link Relative for the

current year
Chain relative for the

previous year
100
Taking the data from illustration 10, we can show the method of conversion as follows:
Year
1990
1991
Price of wheat
100
120
Link relative
100.00
120.00
1992
125
104.167
= 125
1993
130
104.00
= 130
1944
140
107.692
= 140
1955
150
107.14
=150
126
Chain relative
100
= 120
5.8
Base Shifting, Spicing and Deflating the Index Numbers
Base Shifting:
Sometimes it becomes necessary to shift the base from one period to another. This becomes necessary
either because the previous base has become too old and has become useless for comparison purposes or
because comparison has to be made with another series of index numbers having different base period.
The following formula must be used in this method of base shifting:
Index Number (based on New Base Year)
Illustration: 11
Shift the base of the following series to 1997.
Year
Index No.
1995
125
1996
155
1997
185
1998
220
1999
265
2000
320
Solution:
To shift the base at 1997, we multiply every index number by 100/185.
Year
Index No.
1995
67.6
1996
83.8
1997
100
1998
118.9
1999
143.2
2000
173.0
Splicing Two Index Number Series:

The statistical method connects an old index number series with a revised series in order to make the series
continuous is called splicing. The articles which are included in an index number may become out of fashion
or go out of the market. New articles come into the market, for which relative importance may also change.
So it is necessary to include the articles in the index number. The old series of index number is discontinued
and we must construct a new series and must take the year of discontinuation as the first base.
Thus we connect the new set of index with the old discontinued one. The formula is:
Illustration: 12
Two sets of Indices, one with 1986 as base and the other with 1994 as base are given below :
(a) Year
1986
1987
1988
1989
1990
1991
1992
1993
1994
Index Numbers
100
110
120
190
300
330
360
390
400
(b) Year
1994
1995
1996
1997
1998
1999
2000
Index Number
100
105
90
95
102
110
96
The Index (a) with 1986 base was discontinued in 1994. You are required to splice the second index
number (b) with 1994 base to the first index number.
127
Solution:
1986
Year
1986
1987
1988
1989
1990
1991
1992
1993
1994
Splicing of1994
Index Numbers
Index Number (a)

with 1976 as base
100
110
120
190
300
330
360
390
400
Index Number (b)

with 1984 as base
Index Number (b) spliced to

(a) with 1976 as base
100
1995
105
1996
90
1997
95
1998
102
1999
110
2000
96
Deflating:
Deflating is the process of making allowances for the effect of changing price levels. With increasing price
levels, the purchasing power of money is reduced. As a result, the real wage figures are reduced and the real
wages become less than the money wages. To get the real wage figure, the money wage figure may be
reduced to the extent the price level has raised. The process of calculating the real wages by applying index
numbers to the money wages so as to allow for the change in the price level is called deflating. Thus,
deflating is the process by which a series of money wages or incomes can be corrected for price changes to
find out the level of real wages or incomes. This is done with the help of the following formula:
Real Wage
Real Wage Index
Illustration: 13
Given the following data:
Year
1991
1992
1993
1994
1995
1996
Weekly take-home pay (wages)

109.5
112.2
116.4
125.08
135.4
138.1
(1) What was the real average weekly wage for each year?
128
Consumer Price Index

112.8
118.2
127.4
138.2
143.5
149.8
(2) In which year did the employee have the greatest buying power?
(3) What percentage increase in the weekly wages for the year 1996 is required, if any, to provide the same
buying power that the employees enjoyed in the year in which they had the highest real wages?
Solution:
Calculation of Real Wages
(1)
Year
Weekly take-home
pay (Rs.)
Consumer price
Index
1991
109.5
112.8
1992
112.2
118.2
1993
116.4
127.4
1994
125.08
138.2
1995
135.4
143.5
1996
138.1
149.8
Real Wages
Real average weekly wage can be obtained by the formula:

Real Wage
(2)
(3)
The employee had the greatest buying power in 1991 as the real wage was maximum in 1991.
Absolute difference = 97.07 - 92.19 = + 4.88
5.9
Limitations of Index Numbers
Even though index numbers are very important in business and economic activities, they have their own
limitations; they are:
1. The index numbers are only approximate indicators because their construction based on the sample data,
and may not exactly represent the true changes in relative level of a phenomenon.
2. An index number does not take into account the quality of items.
3. Likelihood of error is possible at each stage of construction of index numbers, viz., (i) selection of
commodities, (ii) selection of the base period, (iii) collection of data - prices and quantities of commodities,
(iv) choice of formula - the procedure of weight age to be given.
4. Index number is an average and as such it suffers from all the limitations of an average.
5. Consumption is the result of taste, custom, attitude, etc., which are dynamic.
6. There is no unique index number that is acceptable to all.
7. There may be possibility of manipulation of the base year, price, commodities and quantity quotations in
order to get the required results by the selfish persons.
129
5.10 Summary
Index numbers help to get an idea of the present day situation with regard to changes in production,
consumption, exports and imports, national income, business level, cost of living, the price of a particular
commodity or a group of commodities, Industrial or agricultural production, stocks and shares, sales and
profits of a business house, volume of trade, factory production, wage structure of workers in various
sectors, bank deposits, foreign exchange reserves, and so on.
Index numbers help in formulating policies and decision making, reveal trends and tendencies, measure the
purchasing power of money, aid in deflation, and work like economic barometers. Index numbers may be
divided into three categories price index, quantity index, and value index.
Construction of index numbers requires a careful study of some aspects like purpose of index numbers,
selection of base period, selection of commodities or items, selection of weights, collection of data, selection
of average, price collection, and selection of appropriate formula to construct the index numbers. Methods
of constructing index numbers can broadly be divided into two classes namely: un-weighted index numbers,
and weighted index numbers. In case of un-weighted indices, weights are not assigned, whereas in the
weighted indices weights are assigned to the various items. Each of these types may be further classified
under two heads: aggregate of prices method, and average of price relatives method.
Even though index numbers are very important in business and economic activities, they have their own
limitations; like index numbers are based on the sample data, and may not exactly represent the true changes
in relative level of a phenomena, error is possible at each stage of construction of index numbers, no unique
index number is acceptable to all, and there is a possibility of manipulation of the base year, price, commodities
and quantity quotations.
To conclude we can say that Index numbers are today one of the most widely used statistical devices. They
are used to take the pulse of the economy and they have come to be used as indicators of inflationary or
deflationary tendencies
5.11 Key Words
Index Number: A ratio of the variable value at the current level to the base level, i.e. the ratio of
variable change over a period.
Un-weighted Aggregates Index: Using all the values or data collected for study and allocating
same importance to all the values.
Un-weighted Average of Relatives Method: Working out the index number by dividing the
present or current level of the variable to its base value, multiplied by 100 and then dividing the
summation of percentage values by the number of products to result in average value.
Weighted Aggregates Index: Using all the values considered, but assigning importance or weights
to individual ratios.
Weighted Average of Relatives Method: Constructing the index number by allotting weight
ages to values of each element in the composite.
Consumer Price Index: Indicates the variations in the prices of a given set of consumer items
prepared either at regional level or at National level.
Fixed Base Method: When weight or importance to the variable is allotted based on a given
specified fixed period.
Price index: Compares levels of prices from one period (the current) to another (the base) period.
Quantity index: An index comparing the quantity of the variable during a given period by time.
130

1.
Index numbers are economic barometers. Explain this statement and mention what precautions
should be taken in making use of any published index numbers.
2.
Define Index Number and mention its uses.
3.
Define index numbers. Distinguish between fixed base and chain base method of constructing index
numbers.
4.
What are index numbers? What purpose do they serve? Discuss the various problems faced in the
construction of index numbers.
5.
State and explain Fishers Ideal Formula for Price Index Number and why is it called Ideal?
6.
Write explanatory notes on the following:

(a) Deflating;
(b) Splicing
(c) Base Shifting;
7.
Discuss briefly the uses and limitations of index numbers of prices.
8.
Compute price index for the following data by (i) simple aggregative method, and (ii) average of
price relative method by using arithmetic mean.
Commodities
Price 2003 (Rs.)
Prices 2004 (Rs.)
A
20
25
B
30
30
C
10
15
D
25
35
E
40
45
F
50
55
Ans. (i) 117.14, (ii) 122.9 (by A. M.)

9.
Calculate weighted aggregative price index number taking 2001 as base, from the following data:
(Commodity)
Wheat
Rice
Gram
Pulses
(Quantity
consumed)
4 Qtl.
1 Qtl.
1 Qtl.
2 Qtl.
Base year 2001

80
120
100
200
Price per Unit

Current year 2004
200
250
150
300
Ans. Weighted Index Number = 191.49

10.
From the data given in the following table, calculate consumers price index numbers for the year
2004 taking 2003 as base using (i) simple average, and (ii) weighted average of price relatives:
Items
Wheat
Milk
Egg
Sugar
Shoes
Unit
Kg.
Liter
Dozen
Kg.
Pair
Price in Rs.
2003
0.50
0.60
2.00
1.80
8.00
2004
0.75
0.75
2.40
2.10
10.00
Weight
2
5
4
8
1
Ans. (i) 127.34, (ii) 123.3

11.
Find out the index number for the year 2004 from the following data using the weighted average of
price relatives method:
131
Commodity
Price
Weight
2000
50
30
20
60
20
4
3
5
3
5
Wheat
Milk
Egg
Sugar
Shoes
2004
100
90
10
90
120
Ans.: 270
12.
From the following data, calculate Fishers Ideal Index:

Items
A
B
C
D
13.
Quantity used
2003
2004
5
5
10
11
6
6
4
8
Ans.: 148.78
Given below are two series of index numbers, one based on 1997 and the other on 2000. Splice
the new series on 1997 base:
Year
Old Series (A)
New Series (B)
14.
Price per unit (Rs.)

2003
2004
9.25
15.00
8.00
12.00
4.00
5.00
1.00
1.25
1997
100
-
1998
110
-
1999
125
-
2000 2001 2002 2003 2004

150
100
105
120
130
150
Ans.: 100, 110, 125, 150, 157.5, 180, 195, 225
From the data given below construct index number of quantities, and of prices for 1970 with 1966
as base using (i) Laspeyres formula, (ii) Paasches formula, and (iii) Fishers Ideal formula.
1966
Commodity
15.
Price (Rs.)
Quantity
Price (Rs.)
Quantity
(Units)
(Units)
5.20
100
6
150
A
4.00
80
5
100
B
2.50
60
5
72
C
12.00
30
9
33
D
Ans. (130.07, 131.02, 130.54) and for price (116.29, 117.14, 116.7)
From the following data calculate index numbers of real wages with 1999 as the base:
Year
Average Wages (Rs.)
Consumer Price Index
16.
1970
1996
2400
100
1997
2640
120
1998
2860
130
1999
3000
150
2000
3420
190
2001
4000
200
2002
4200
210
Ans. 120, 110, 110, 100, 90, 100, 100

From the data given below, calculate the price index by Fishers ideal formula and then verify that
Fishers ideal formula satisfies both time reversal test and factor reversal test.
Commodity
A
B
C
Base year
Price (Rs.)
Quantity ('000
tonnes)
56
71
32
107
41
62
Current year
Price (Rs.)
Quantity ('000
tonnes)
50
26
30
83
28
48
Ans. (84.92)
132
5.13 References
1.
Richard I. Levin and David S. Rubin, Statistics for Management
2.
Gupta, S. P., Statistical Methods
3.
Yadav, Jain, Mittal, Statistical Methods.
4.
Nagar, K. N., Statistical Methods.
5.
Gupta, C.B. and Gupta, Vijay, An Introduction to Statistical Methods.
133
Unit - 6 : Probability and Probability Distributions

Unit Structure:
6.0
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.8
6.9
6.10
6.11
6.12
6.13
6.14
6.15
Objectives
Introduction
Approaches
Addition Law
Multiplication Law
Conditional Probability
Probability Function
Bernoulli Distribution
Poisson Distribution
Normal Distribution
Main Characteristics of Normal Distribution
Importance of Normal Distribution
Summary
Key Words
References
6.0
Objectives
After studying this unit, you should be able to understand
Probability concepts and rules
Probability distribution Concept
Binomial distribution and its uses
Poisson distribution and its uses
Normal distribution and its uses
6.1
Introduction
Probability has a very old history, it was originated in the games of chance related to gambling. For instance,
throwing of dice or coin and drawing cards from a pack. Jerome Cardan (1501~1576), an Italian
mathematician was the first man to write a book on the subject Book on Games of chance which was
published in 1663 after his death. The probability formulae and techniques were developed by Jacob
Bernoulli(1654-1705), De Moivre (1667-1754), Thomas Bayes(1702-1761) and Joseph Lagrange(17361813). Pierre Simon, Laplace in the nineteenth century unified all these early ideas and compiled the first
general theory of probability.
In the beginning, the probability theory was successfully applied at the gambling tables. But after some time,
it was applied in the solution of social, political, economic and business problems. In fact, it has become a
part of our everyday lives. We face uncertainty in personal and management decisions and use probability
theory. Probability constitutes the foundation of statistical theory.
134
6.2
Approaches
There are mainly three approaches to probability

i.
Classical approach
ii.
Empirical approach
iii.
Axiomatic approach
Few terms can be defined / explained with reference to simple experiments relating to tossing of coins,
throwing of a diee or drawing cards from a pack of cards.
Random Experiment
An experiment can be considered as a random experiment if all the possible outcomes are known in advance
and none of the outcomes can be predicted with certainty. e.g. throwing a dice, tossing a coin etc.
Trial & Event
When a random experiment is performed, it is called a trial and outcome or combinations of outcomes are
termed as events. For example
i.
When a coin is tossed repeatedly, the result is not unique. We may get any of the two faces; head or
tail. Thus, throwing a coin is a random experiment and getting of a head or tail is an event.
ii.
In the similar manner, when a dice is thrown, it is called a random experiment. Getting any of the
faces 1, 2, 3, 4, 5 or 6 is an event. Getting an odd no. or an even no., getting no. greater than 3 or
lower than five, these are called events.
iii.
Similarly, drawing of two balls from an urn containing a red balls and b white balls is a trial and
getting of both red balls, or both white balls, or one red and one white ball are events.
Exhaustive Cases
When a random experiment is done, there are some outcomes; the total numbers of possible outcomes are
called exhaustive cases for the experiment. For e.g. when a coin is tossed, we can get head (H) or tail (T).
Hence exhaustive no. of cases is 2 (i.e. H,T) If two coins are tossed, the various possibilities are HH, HT,
TH, TT (number of exhaustive cases are four) where HT means, Head on first coin and Tail on second coin,
and TH means, Tail on first coin and Head on second coin and so on.
In case of toss of three coins, number of outcomes is
= (H,T) x (H,T) x(H,T)
= (HH,HT,TH,TT) x (H,T)
= (HHH, HHT, HTH, HTT, THH, THT, TTH, TTT)
No. of possible outcomes is 8 = 23. In general, in a throw of n coins, the exhaustive no. of cases is 2n.
In a throw of a die, exhaustive number of cases is 6, since we can get any one of the six faces marked 1, 2,
3, 4, 5 and 6. If two dice are thrown, the possible outcomes are
(1,1)
(2,1)
(3,1)
(1,2)
(2,2)
(3,2)
(1,3)
(2,3)
(3,3)
(1,4)
(2,4)
(3,4)
135
(1,5)
(2,5)
(3,5)
(1,6)
(2,6)
(3,6)
(4,1)
(5,1)
(6,1)
(4,2)
(4,2)
(4,2)
(4,3)
(4,3)
(4,3)
(4,4)
(4,4)
(4,4)
(4,5)
(4,5)
(4,5)
(4,6)
(4,6)
(4,6)
i.e. total no. is 36, where (i,j) means number i on the first die and j on the second die, i and j both taking the
values from 1 to 6. In the case of throw of two dice, no. of possible outcome = 62 = 36 and in the case of
throw of three dice, no. of possible outcome = 63 = 216; in the case of throw of n dice, no. of possible
outcome = 6n
Favourable cases or events
The number of outcomes of a random experiment which result in the happening of an event are named
as the cases favourable to the event.
i. When a toss of two coins takes place, the no. of cases favourable to the event exactly one head is
two (i.e. TH or HT) and for getting two heads is one (i.e. HH)
ii. When a card is drawn from a pack of cards, the no. of cases favourable to the event getting a
diamond are 13 and getting an ace of spade is one.
Mutually Exclusive events or cases
Two or more events are considered as mutually exclusive if the happening of any one of them excludes the
happening of all others in the same experiment. For example, in toss of a coin, the events head and tail
are mutually exclusive because if head comes, we cant get tail and if tail comes we cant get head. Similarly,
in the throw of a die, the six faces numbered 1, 2, 3, 4, 5 and 6 are mutually exclusive. Thus, events are said
to be mutually exclusive of no two or more of them can happen simultaneously.
Classical/priori Probability
It is the oldest and simplest approach. Under this approach, there is no need to physically perform the
experiment. The basic assumption is that the outcomes of a random experiment are equally likely. e.g. in a
throw of a dice, occurrence of 1,2,3,4,5,6 are equally likely event.
If a random experiment results in N exhaustive, mutually exclusive and equally likely outcomes out of which
m are favourable to the happening of an event X then the probability of occurrence of X i.e. P(X) is given by
Favourable cases
P(X) = Exhaustive cases
m
N
Example 1. A bag containing 10 green and 20 red balls. A ball is drawn at random. What is the
probability that it is green.
Sol. Total number of balls in the bag = 10+20 = 30
Number of green balls = 10
Probability of getting a green ball =
Favourable cases
Exhaustive cases
136
10
1
=
30
3
Empirical Probability
The classical definition is difficult to apply as soon as we move from the field of coins, cards, dice and other
games of chance. It may not explain the actual results in certain cases e.g if a coin is tossed 20 times , we
may get 14 heads and 6 tails. The probability of head is thus 0.7 and tail is 0.3. However, if experiment is
carried out large number of times, we should expect approximately equal number of heads and tails.
If an experiment is performed repeatedly under essentially homogeneous and identical conditions then the
limiting value of the ratio of the number of times the event occurs to the number of trials, as the number of
trials become indefinitely large is known as the probability of happening of the event.
P(X) = lim
m
N
N
Axiomatic Approach
The axiomatic Probability theory is an attempt at constructing a theory of probability which is free from
inadequacies of both the classical and empirical approaches. It plays an important role in rendering a
reasonable amount of comprehensibility and tractability to the understanding of chance phenomenon atleast
in the initial stages of any scientific inquiry into their structure and composition where other approaches are
less comprehensible and tractable.
6.3
Addition Law
The probability of occurrence of either event A or event B of two mutually exclusive events is equal to the
sum of their individual probability.
Mathematically, we can represent as
P(A B) = P(A) + P (B)
Proof :- If an event A can happen in a1 ways and B in a2 ways then
The number of ways in which either event can happen in a1 + a2 ways.
Total number of possibilities is n.
Then by definition , the probability of either the first or second event happening is
a1
a1 + a 2
a2
=
+
n
n
n
But
And
a1
n
= P(A)
a2
= P(B)
n
Hence P(A B) = P(A) + P (B)

137
The theorem can be extended to three or more mutually exclusive events, thus
P(A B C) = P(A) + P (B) + P(C)
Example 2. A deck of 52 cards, one card is drawn. What is the probability that it is either a king or a
queen?
Sol. There is four kings and four queens in a pack of 52 cards.
The probability of drawing a card that is king =
4
52
The probability of drawing a card that is queen =
4
52
Since the events are mutually exclusive, the probability that the card drawn is either a king or
4
4
2
a queen =
+
=
52
52 13
If two events A & B are not mutually exclusive (joint events) then the addition law can be stated as
follows
The probability of the occurrence of either event A or event B or both is equal to the probability that event
A occurs, plus the probability that event B occurs minus the probability that both events occur. I can be
shown as
P(A B) = P(A) + P (B) P (A B)
Proof.
P(A B) = n(A B)/n(U)
Where n(A B) - No. of elements belonging to (A B)
And n(U) total no. of elements in universal set U.
A B
P(A B)=
n(A)+ n(B)n(A B)
n(U)
( )
n(U )
( )
n(U)
B)
n(U )
= P(A) + P(B) - P(A B)
138
Example 3. The managing committee of Residents Welfare Association formed a sub-committee of 5

persons to look into electricity problem. Profiles of the 5 persons are
Male age
40
Male age
43
Female age
38
Female age
27
Male age
65
If a chairperson has to be selected from this, what is the probability that he would be either female or
over 32 years.
Sol. P (female or over 32) = P(female) + P(over 32) P(female and over 32)
=
2
4 1
+
=1
5
5 5
Example 4. What is the probability of picking a card that was a heart or a spade.
Sol. Using the addition rule,
P(heart or spade) = P(heart) + P(spade) P(heart and spade)
=
6.4
13
13 1
1
+
=
52
52 2
2
Multiplication Law
It states that if two events A and B are independent , the probability that they both will occur is equal to
the product of their individual probability.
P(A and B) = P(A) X P(B)
It can be extended to three or more independent events.
P(A,B and C) = P(A) X P(B) X P(C)
Proof :- If an event A can happen in n1 ways of which a1 are successful and B in n2 ways of which a2 are
successful then
The number of successful happening in both cases is a1x a2 .
Total number of possibilities is n1 x n2.
Then by definition , the probability of the occurrence of both events is
a1 x a2
a1
a2
=
x
n1 x n2
n
n
But
a1
= P(A)
n
and
a2
= P(B)
n
139
Hence P(A and B) = P(A) x P (B)

Example 5. In order to marry with a girl, a man wants these qualities
White complexion the probability of getting this is one in fifty.
Etiquettes the probability is one in hundred.
Dowry the probability of getting this is one in Twenty.
Calculate the probability of his getting married to such a girl when the possession of these three attributes
is independent.
Sol.
1
= 0.05
20
Probability of a girl with white complexion =
Probability of a girl with handsome dowry =
Probability of a girl with etiquettes =
1
= 0.02
50
1
= 0.01
100
The probability of simultaneously occurrence of all these qualities = 0.05 x 0.02 x 0.01
= 0.00001
6.5
Conditional Probability
The multiplication theorem described above is not applicable in case of dependent events. Two events A
and B are said to be dependent when B can occur only when A is known to have occurred. The probability
attached to such an event is called conditional probability and is denoted by P(A/B) i.e. probability of A
given that B has occurred.
Similary ,
P ( A / B ) P ( A B ) / P ( B );
P( B) 0
P ( B / A) P ( A B ) / P ( A);
P ( A) 0
Symbolically , we write
or
6.6
P( A B) P( A / B) P( B)
P ( B A) P ( B / A) P ( A)
Probability Function
If the function permits us to compute the probability for any event that is defined in terms of value of the
random variable, then the function is called a probability function. Just as there are discrete and continuous
random variables, so there are discrete and continuous probability functions. Different types of probability
functions are shown:
140
PROBABILTY FUNCTION
DISCRETE PROBABILTY
FUNCTION
PROBABILTY
MASS
FUNCTION
CONTINUOUS PROBABILTY
FUNCTION
CUMULATIVE
MASS
FUNCTION
PROBABILTY
DENSITY
FUNCTION
CUMULATIVE
DENSITY
FUNCTION
Discrete Probability Function

A probability function for a discrete random variable is called a discrete probability function since the
domain of the function is discrete.
Probability Mass Function
A probability function that specifies the probability that any single value of discrete random variable will
occur is called a probability mass function. If f(x) is the probability mass function of the random variable X,
then f(x) = P(X=x) has the following properties
(i)
f(x) 0 for all values of X; and
(ii)
f(x) = 1
Cumulative Mass Function
If X is a discrete random variable with p.m.f f(x), its cumulative mass function
(c.m.f.) specifies the probability that an observed value of X will be no greater than X. i.e. if
F(x) is a c.m.f. and f(x) is a p.m.f. then F(x) = P( X x ).
Continuous Probability Function
A probability function for a continuous random variable is called a continuous probability function since the
domain of the function is continuous.
Probability Density function
For a continuous random variable, the corresponding function f(x) is called a probability density function
(p.d.f.). Unlike a p.m.f., a p.d.f. doesnt specify probabilities for specific individual value of the random
variable.
Cumulative Density function
Corresponding to the cumulative mass function of a discrete random variable, the cumulative densityfunction
(c.d.f.) of a continuous random variable specifies the probability that an observed value of X will be no
greater than x.
141
6.7
Bernoulli Distribution
BERNOULLI TRIALS A random variable X which takes two values 0 & 1 with probabilities p and q
respectively
It was discovered by James Bernoulli (1654 1705). Consider a set of n Bernoulli trials. Following are
conditions for Bernoulli trials
a) An experiment is performed for a fixed number of trials.
b) In each trial, there are two possible outcomes of the experiment i.e. success or failure.
c) The probability of a success(p) remains constant from trial to trial. If the probability of success is
not the same in each trial, we will not have binomial distribution.
d) The trials are statistically independents i.e. the outcomes of any trial do not affect the outcomes of
subsequent trials.
A random variable X is said to be followed binomial probability distribution if it assumes only non negative
values and its probability distribution is given by
P (X = x) = p (x) = { ncx px qn-x , x= 0, , 2 _ _ _ _ , n
A random variable which satisfies conditions of Bernoulli trials, can be represented by Binomial Probability
Distribution.
Binomial Distribution is denoted as
X ~ B (n,p) where q = 1 p
n,p are the parameters of binomial distribution and n is the degree of binomial distribution.
Some formulas are
Mean of Binomial Distribution = np
Variance of Binomial distribution = npq

Standard Deviation of Binomial Distribution =
npq
Example 6. Find n,p and q when the mean of a binomial distribution is 40 and standard
deviation is 6.
Sol. Mean of a binomial distribution = np
Standard deviation of a binomial distribution =
npq = 6 (given)
..........(1)
np = 40 (given)
..........(2)
npq = 6 gives
npq = 36;
.............(3)
142
npq
Putting (2) in (3)

40. q = 36, q = 36/40
q = 0.9, putting value in p = 1 q,
gives p = 0.1
np = 40
40
0.1
= 400
Ans. n=400, p = 0.1, q = 0.9
Example 7. Suppose that the half of the population of town are consumers of rice. One hundred investigators
are given the duty to search for the truth. Each investor investigates 10 individuals. How many investigators
do you expect to report that three or less of the people interviewed in there sample are consumers of rice?
Sol. Given n = 10
P=
1
2
Probability that three people or less consume rice P (X 3)
= P [X = 0] + P [X = 1] + P [X = 2] + P [X = 3]
= 10C0 p0 q10 + 10C1 p1q9 + 10C2 p2q8 + 10C3 p3 q7
1
1
1
1
= ( )10 + 10 ( )10 + 45 ( )10 + 120 ( )10
2
2
2
2
1
= ( )10 (1 + 10 + 45 + 120)
2
176
=
1024
Therefore the no. of investigators to report that three people or less consume rice is
given by NP where N is total number of trials.
176
=
x 100
1024
= 17.2 = 17 (approx.)
Example 8. In an industry, the workers have a 20% chance of suffering from an occupational disease.
What is the probability that out of six workers, 4 or more will contract the disease?
Sol. The probability that a worker is suffering from the disease (p) = 20 / 100,
The probability that a worker is not suffering from the disease
143
q = 1-p =1- 1/5 = 4/5

Probability that 4 people or more; i.e. 4, 5 or 6 will contract disease P (X 4)
= P [X = 4] + P [X = 5] + P [X = 6]
1
4
1
4
1
4
= 6C4 ( )4 ( )2 + 6C5 ( )5 ( )1 + 6C6 ( )6 ( )0
5
5
5
5
5
5
1
4
1
4
1
4
= 15( )4 ( )2+ 6 ( )5 ( )1 + ( )6 ( )0
5
5
5
5
5
5
16
4
1
= 15 x
+6x
+
15625
15625 15625
= 0.01696
6.8
Poisson distribution was developed by the French mathematician and physicist Simeon Denis Poisson
(1781 1840). This distribution is frequently used in context of Operation Research. This distribution plays
an important role in Inventory control problems, Queuing theory and also in Risk models. Unlike binomial
distribution, Poisson distribution can not be deduced on purely theoretical grounds based on the experiment
conditions. In fact, it must be based on empirical results of past experiments relating to the problem under
study.
Poisson distribution is a limiting case of a binomial distribution under the following conditions
i.
n, the no. of trials is indefinitely large i.e. n
[ n is positive integer]
ii.
p, the constant probability of success for each trial is indefinitely small i.e. p
iii.
np = m is finite where m is a positive real number
Therefore, lim. b(x; n, p) =

n
Definition
A random variable X is said to follow a Poisson distribution if it assumes only nonnegative values and its probability distribution is given by
P(x,m) = P (X=x) =
= 0;
; x = 0, 1, 2, -------------,
otherwise
Here, m is known as the parameter of the distribution and is equal to np.

Following are conditions for Poisson Distribution
a) It is a discrete probability distribution with
i) Mean = m
ii) Variance = m
144
b) Total number of terms are .

c) Probability of success is constant from trial to trial for any given specific interval size.
Example 9. The standard deviation of Poisson variable X is 2, Find the probability that X is strictly
positive.
Sol. We know that for Poisson distribution with parameter m,
Variance = m = (2)2 = 2
[Because S.D. = 2, given]
Therefore, P(X = r) =
;r = 0,1,2.......
The probability that X is strictly positive is given by

P( X > 0 ) = 1 P ( X = 0 ) = 1 - e-2
Answer is 1 - e-2
Example 10. On the basis of past data and experience, it was found that in a plant, there are on an average
four accidents per month. Find the probability that in a particular month, there will be less than four accidents.
Assuming Poisson distribution (e-4 = 0.0183)
Sol. We suppose that random variable X indicates the no. of accidents in the plant per month, (m is
given as 4 in the usual notation) then by Poissons probability law
P(X = r) =
The probability that no. of accidents in a particular month in the plant is less than 4; is given by
P (X< 4) = P (X = 0) + P (X = 1) + P (X = 2) + P (X = 3)
= e-4 [1 + 4 + 42 / 2! + 43 / 3!]
= e-4 [1 + 4 + 8 + 10.67]
= e-4 x 23.67 = 0.0183 x 23.67 = 0.4332
Example 11. A manufactured product has 2 defects per unit of product inspected. Using Poisson distribution,
calculate the probabilities of finding a product
a) without any defect
b) three defects
c) four defects
Given e-2 = 0.135
Sol.
Average number of defects m = 2
145
P(0) = e-2 = 0.135 (Given)

P(1) = P(0) x m
= 0.135 x 2
= 0.27
P(2) = P(1) x
m
2
= 0.27 x
2
2
= 0.27
P(3) = P(2) x
= 0.27 x
m
3
2
3
= 0.18
P(4) = P(3) x
m
4
= 0.18 x
2
4
= 0.09
Hence the probability that a product has no defect is 0.135, product has 3 defects is 0.18 and product
has 4 defects is 0.09.
6.9
Normal Distribution
The normal distribution was discovered in 1733 by English mathematician De-Moivre who obtained this
continuous distribution as a limiting case of the binomial distribution and applied it to problems arising in the
game of chance throughout the eighteenth and nineteenth centuries, various efforts were made to establish
the normal model as the underlying law ruling all continuous random variables. Thus the name normal.
These efforts however failed because of false premises. The normal model has nevertheless become the
most important probability model in statistical analysis.
Definition
A random variable X is said to follow a normal distribution with parameters (called mean) and variance
if its density function is given by the probability law
f(x; , ) =
or f(x; , ) =
where - < x < +
- < < +
>0
146
Note
i.
A random variable x with mean variance 2 and following the normal law is
expressed by X ~ N (, 2)
ii.
If z = x - /
(z) = 1/ 2 . exp.[ -z2/2]
Mean of z = E (z) = E [(x- ) / ]
= E(x) - /
=- /
=0
Variance of z = 1
Z is known as standard normal variate
- < z < +
6.10 Main Characteristics of Normal Distribution

Normal probability curve
x =
Fig. 6.1
The normal probability curve with mean and standard deviation is given by the equation
F(x) =
- < x < +
and has the following characteristics

i. The curve is bell shaped,unimodal and symmetrical about the line x = .
ii. Mean, median and mode of the distribution coincide.
iii. Since f(x) being the probability, can never be negative, no portion of the curve lies below
the x-axis.
147
iv. Linear combination of independent normal variates is also a normal variate.

v. X-axis is an asymptote to the curve.
vi. The points of inflexion of the curve are given by [x = , f(x) = 1/ 2
. e- 1/ 2 ]
vii. Area Property
P( < x < + )
= 0.6826
P( 2 < x < + 2) = 0.9544

P( 3 < x < + 3) = 0.9973
(The curve is given below, Fig. 6.2)
-3 -2 -
-3
-2
x = + +2 +3
-1
z =0
Fig. 6.2
x1
x1
P ( x / x1 ) f ( x )dx 1 / / 2 fe
( x u)2
dx
2 2
The probability that random value of x will lie between x = and x = x1 is given by
Put ((x ) / ) = z
When x = , z = 0 and
When x = x1 , z = ((x1 ) / ) = z1 (say)
Therefore, P( < x < x1) = P (o < z < z1)
z1
z1
Z2
= 1/ 2 e 2 dz = (z) dz
0
Where (z) = 1/ 2 . e - -
Z2
is the probability function of standard normal variate. The definite integral
2
z1
148
(z) dz is known as normal probability integral and gives the

0
area under standard normal curve between the coordinates at z = 0, and z = z1
In particular, the probability that a random value of x lies in the interval (, +) is given by
+
P ( < x < +) = f(x) dx
1
P (-1 < z < 1)
=
(z) dz
-1
1
=
2 (z) dz
0
1
1
= 2/ 2 e 2 dz = 2 x 0.3413
0
= 0.6826
Similarly,
P (2 < x < +2)
= P (-2 < z < 2)
2
= 2 (z) dz
0
= 2 x 0.4772 = 0.9544
and
3
P (3 < x < +3) = P (3 < z < 3) =
= 2 x 0.49865
(z) dz
-3
0.9973
Thus, the probability that a normal variate x lies outside the range 3 is given by
P [(x- ) > 3] = P (z > 3)

= 1 P (-3 z 3)
= 1 -0.9973
= 0.0027
149
6.11 Importance of Normal Distribution

Normal distribution plays a very important role in statistical theory because of the following reasons
ii
i. Most of the distribution occurring in practice e.g. Binomial, Poisson etc. can be approximated by
normal distribution.
ii. Even if a variable is not normally distributed it can sometimes be brought to normal form by simple
transformation of variable.
i.
P (-3 < x < +3) = 0.9973

P (-3 < z < 3) = 0.9973
P (z < 3) = 0.9973
P (z > 3) = 0.0027
This property of the normal distribution forms the basis of entire large sample theory.
iv. Theory of normal curves can be applied to the graduation of the curves which are not normal.
v. Normal distribution finds large applications in Statistical Quality Control in industry for setting control
limits
vi. Many of the distributions of sample statistics (e.g. the distribution of sample mean, sample, variance
etc.) tends to normality for large samples and as such they can best be studied with the help of
normal curves.
vii. The entire theory of small sample tests is based on the fundamental assumption that the parent
populations from which the samples have been drawn follow normal distribution.
Example 12. The annual demand for one type of fertilizer is normally distributed with a mean of 120 tonnes
and standard deviation of 16 tonnes. If a wholesale distributor orders only once a year, what quantity should
be ordered to ensure that there is only 5% chances of running short of quantity?
Sol. We suppose that the annual demand (in tonnes) be denoted by the random variable X.
Therefore, z = (X-120) / 16
The desired area of 0.05 can be shown in the figure below. Since the area between the mean and the given
value of X is 0.45, therefore from the table we get this area of 0.45 corresponding to z = 1.64
0.45
120
Shaded Area
150
By substituting value of z = 1.64 in standardized normal variate, we will get

1.64 = (X 120) / 16
X = 120 + (1.64) (16) = 146.24
We can take the order value 146.24 as 147 by rounding if the ordered value is a whole unit.
Example 13. The income distribution of officers of a certain company was normally distributed with a
mean income per month Rs. 15,000 and standard deviation Rs. 5,000. If a there were 242 officers,
drawing salary above Rs. 18,500, how many officers were there in the company?
[The area under standard normal curve between 0 and 0.7 is 0.2580] Given
Sol. We suppose that the income of the officers (in Rs.) be denoted by the random variable X.
Therefore, X N( ,) where = Rs. 15000, and = Rs. 5000
Let N be the no. of officers in the company, then the no. of officers in the company drawing salary above
Rs. 18500 is
N. P (X > 18500) = 242 (given) (1)
When X = 18,500, Z = (X ) / = (18500 15000) / 5000 = 0.7
Therefore, P(X > 18500) = P (Z> 0.7)

= 0.5 P(0 Z 0.7)
= 0.5 0.2580
= 0.242
Substituting value of P(X > 18500) in equation (1)
We get from eq. N. P (X > 18500) = 242
N. 0.242 = 242
N =1000
Example 14. There are 1000 soldiers in a regiment, it is assumed that the average height of soldiers is
68.22 inches, variance being 10.8 inch2. How many soldiers in the regiment would you expect to be
i)
Over six feet tall
ii)
less than 5.5 feet.
It is assumed that height is normally distributed.

Sol. We suppose that the height of soldiers (in inches) be denoted by the random variable X. Then, we
are given
Mean = = 68.22
and
= 10.8
A soldier will be six feet if X is greater than 72 (because X is height in inches).

When X = 72, Z = (X ) / = (72 68.22) /
10.8 = 3.78 / 3.286 = 1.15
i) The probability that a soldier is over six feet (= 72"), is given by

151
P(X > 72) = (Z> 1.15) = 0.5 P(0 Z 1.15)

= 0.5 0.3749 = 0.1251
(From Normal tables)
Hence, in a regiment of 1000 soldiers, the no. of soldiers over six feet is
1000 x 0.1251 = 125.1 = 125 (rounded)
ii)
The probability that a soldier is below 5.5(= 66"), is given by
P(X < 66) = P(Z < (66 68.22) /
10.8 ) = P(Z < ( 2.22) / 3.286 )
= P(Z< -0.6756) = P( Z > 0.6756)

= 0.5 P(0 < Z< 0.6756) = 0.5 - 0.2501
(From Normal tables)
= 0.2499 (approx.)
Hence, the no. of soldiers over 5.5 in a regiment of 1000 soldiers is
1000 x 0.2499 = 249.9 = 250 (rounded)
6.12 Summary
This unit provides you a detailed description about probability, various probability laws and probability
distribution. Under given conditions, observed frequency distributions can be approximated by well known
theoretical distributions. The theoretical distributions indicate as what to expect if a random variable behaves
as we assume it does. Probability distributions fall in two categories:
Discrete
Continuous
Most important discrete distributions are Binomial and Poisson. Similarly important continuous distribution
is Normal distribution.
6.13 Key Words
Random Experiment An experiment can be considered as a random experiment if when conducted

rapidly under essentially homogeneous conditions, the result is not unique but may be any one of the
various possible outcomes.
Trial & event When a random experiment is performed, it is called a trial and outcome or
combinations of outcomes are termed as events.
Exhaustive Cases When a random experiment is done, there are some outcomes; the total
numbers of possible outcomes are called exhaustive cases for the experiment.
Mutually Exclusive events or cases Two or more events are considered as mutually exclusive
if the happening of any one of them excludes the happening of all others in the same experiment

1.
Explain in detail
Binomial Distribution
152
Normal Distribution
2.
Explain Probability Function in detail with diagram.
3.
Out of 800 families with 4 children each, what percentage would be expected to have
a)
At least one boy
b)
No girls
c)
Two boys and two girls
d)
At most 2 girls
Assume equal probability for boys and girls.

(Ans. (a) 93.75%
4.
(c).37.5%
(d) 68.75%)
There are 1000 employees, mean of their wages is Rs. 70. Daily wages are distributed normally
distributed around the mean and their Standard deviation is Rs. 5. Find out Number of workers
whose daily wages will be
i.
Between Rs. 70 and 72
ii.
Between Rs. 69 & 72
iii.
More than Rs. 75
iv.
Also find out the lowest daily wages of the 100 highest paid workers
(Ans. i. 155,
5.
(b) 6.25%
ii. 235,
iii. 159, iv. 76.40)
A company is manufacturing electric bulbs, if 5% of the electric bulbs manufactured by the

company are considered to be defective, use Poissons distribution to find the probability that in
a sample of 100 bulbs
i.
None is defective
ii.
5 bulbs will be defective
(e-5 = 0.007)
(Ans. i. 0.007, ii. 0.1823)

6.
If on an average 8 ships out of 10 arrive safely at a port, find the mean and standard deviation of
the number of ships arriving safely out of a total of 1600 ships.
(Ans. 1280 ,16)
7.
A machine is processing screws and screws are being checked for quality by examining number of
defectives in a sample of 6. The following table shows the distribution of 128 samples according to
the number of defective items they contained
No. of defectives
total
No. of samples
19
35
30
23
128
i)
Fit a binomial distribution and find the expected frequencies if the chance of screw being
defective is .
153
ii)
8.
Find the mean and standard deviation of the fitted distribution.
Ten unbiased coins are tossed concurrently. Find the probability of obtaining
i.
Exactly six heads
ii.
No head
(Ans. i. 105/512,
iii
At least 8 heads
iv At least one head

ii. 1/1024,
iii. 7/128,
iv. 1023/1024)
9.
Find the probability that at most 5 defective bolts will be found in a box of 200 bolts if it is
known that 2% of such bolts are expected to be defective. (Given e-4 = 0.0183 )
10.
There are 600 business students in the post-graduate department of a university. The probability for
any student to need a copy of a particular book from the library on any day is 0.05. How many
copies of the book should be kept in library so that the probability may be greater than 0.90 that
none of the students needing a copy from the library has to come back disappointed.
(Ans. 37)
6.15 References
Business Statistics by D.R.Agarwal, Vrinda Publications(P) Ltd.
Business Statistics by G.C.Beri, Tata McGraw Hill Publishing Company Ltd.
Fundamentals of Statistics by S.C.Gupta, Himalaya Publishing House.
Business Statisitcs by S.P.Gupta, Sultan Chand & Sons Educational Publishers, New Delhi.
Statistics for Business and Economics by R.P.Hooda, Macmillan India Ltd.
Research Methodology by C.R.Kothari, New Age International (P) Limited, Publishers.
154
Unit - 7 : Decision Theory

Unit Structure:
7.0
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
7.10
7.11
Objectives
Introduction
One Stage Decision Making Problem
Developing Pay-Off Table
Developing Regret Table
Decision Models
Bayesian Decision Rule: Posterior Analysis
Limitations
Summary
Key Words
References
7.0
Objectives
The concept/steps of decision process
The concept of uncertainty and risk
Various methods in decision under uncertainty
Various methods in decision under risk
The concept and calculation of value of perfect information
The decision tree and its analysis
7.1
Introduction
Every one of us has to make decisions throughout life. What profession to choose? Where and how much
to invest? What to produce and how much to produce? and so on, are some situations where decisions are
to be made? Some of the decisions are really difficult to make because of the complexity of the decision
situation. This is true in business and industrial problems which are becoming more and more complex. The
decision making in business is a very tough task. If someone asks the business executives as to how do they
take decision then it will be very difficult question for them to answer because they hardly follow any
consistent procedure. Sometimes decision based on cost and profit calculations or at other times based on
relative competitiveness. Therefore, we need a decision theory which is defined as a body of methods;
helpful to decision maker to select a course of action amongst the alternative plans of actions open to him.
Decision Theory is concerned with how to assist organizations in making decisions.
It consists of following steps:
Decision making Environment
Objectives of a decision maker
Alternative plans of action
Decision payoff
155
Decision Making Environment

The first step is to know the environment under which a decision is to be taken. Who takes the
decision? This is a relevant question. The decision maker may be an individual or a group of individuals.
Then they take care of decision situations. These situations can be:
Situation of certainty
Situation of Risk
Situation of Uncertainty
Situation of certainty Certainty situation is one where the outcome of a specified decision can be
predetermined with certainty i.e. each action lead to only one outcome. For example if there are three types
of machines say X, Y and Z which can take a given job with same level of required accuracy and if the setup
time for these machines and operating time per piece on them are known and we are to find out on which
machine, an order of given quantity should be processed. Such a situation neither involves competitive nor
probabilistic phenomenon. The various important techniques for taking decisions under condition of certainty
may be:
a) Techniques used in Transportation Problems
b) Simplex method for LPP
c) Activity Analysis
d) Various mathematical calculations and formulae are available for different applications in business.
Situation of Risk Risk situation is one in which decision maker knows the likelihood that each of the
various states of nature will occur i.e. each action will lead to one outcome, each with a known probability.
It is a situation where there are many states and the decision maker knows the probability of occurrence of
each state. Such decision situations are frequent in business and industry for in many business problems the
probabilities of various states are known by determining how frequently they occurred in the past.
Situation of Uncertainty Uncertainty situation is one when the probabilities associated with the state of
nature are unknown. Decision making under uncertainty is more difficult than decision making under risk.
Many of the major decision problems of business involving situation of uncertainty. For example, if an
executive says that the chance that within next six months the project will be completed successfully is 75%
then the executive is making a subjective statement and is based on his past managerial experience. Obviously
under the situation of uncertainty, there is no one best single criterion for taking a decision.
Decision Stage
Following are the decision stages:
(i) Objectives of a decision Maker
The decisions are made because the decision maker wants to achieve some object. A decision maker will
always choose an action that will enable him to achieve his objective. Thus, it is important in the decision
process to define explicitly the objectives involved in the decision making process.
(ii) Alternative Plans of Action
If there is only one course of action, there is no decision problem. In a decision process, there are several
alternative plans of action. Such alternative plans of action may be limited or unlimited. In the market
156
whether to introduce or not a new product, there are only two alternatives. But for advertisement decision,
alternatives would be unlimited. Hence an exhaustive list of all feasible alternative plans of action should be
prepared in advance .The problem of decision making is to pick up the best out of these limited or unlimited
strategies.
(iii) Decision Payoff
Decision payoff is an indication of the effectiveness of the strategies. In general, effectiveness is measured in
terms of money but in many situations it is not possible to give a realistic value of money. Decision payoffs
may be fixed decision payoff or can be random variable. In the former case, it is of deterministic nature and
in the latter case it is probabilistic in nature. Probability payoff is determined by chance and the strategies
chosen.
7.2
One Stage Decision Making Problem
In single stage decision making problems, decisions are taken by considering the pay-offs resulting from
various courses of action and outcomes possible.
Example 1. A book shop sells a tax book for Rs. 100. It purchases that book for Rs. 80 per copy. Since
tax laws change every year, some copies become outdated and can be disposed of for Rs. 30 each. The
annual demand for the book is between 18 and 23 copies according to past experience. Assume that the
order can be placed only once during the year, the problem before the seller is to decide how many copies
of the book should be purchased for the next year.
Sol. Since the annual demand varies between 18 and 23 copies, there are six possible events:
E1 :
E2 :
E3 :
E4 :
E5 :
E6 :
demand for 18 copies

Also, there are six possible courses of action.

A1 :
A2 :
A3 :
A4 :
A5 :
A6 :
buy 18 copies
buy 19 copies
buy 20 copies
buy 21 copies
buy 22 copies
buy 23 copies
After getting the list of possible actions and events, the next step is to construct the pay-off table.
7.3
Developing Pay-Off Table
A pay-off table shows the economics of the given problem. A pay-off is a conditional value may be conditional
cost. Conditional means that it is associated with each course of action given that certain event has occurred.
A pay-off table represents the matrix of conditional values associated with all the possible combinations of
the acts and the events.
Let D demand in units for the book
157
Q quantity decided to be purchased

Then P = 20Q
when DQ
P = 70D 50Q
when DQ
The Pay-off table for the above example given below:
Event Ei
E1 : 18
E2 : 19
E3 : 20
E4 : 21
E5 : 22
E6 : 23
7.4
Action Aj
A1 : 18
360
360
360
360
360
360
A1 : 19
310
380
380
380
380
380
A1 : 20
260
330
400
400
400
400
A1 : 21
210
280
350
420
420
420
A1 : 22
160
230
300
370
440
440
A1 : 23
110
180
250
320
390
460
Developing Regret Table
The resultant outcomes of the various combinations of the actions and events can be expressed alternatively
in terms of regret. It is defined as the amount of pay-off foregone by not adopting the optimal course of
action, which would give the highest pay-off for each possible event. For instance, if the demand is 18
copies then the optimal act is to buy 18 copies then regret is zero. Regret table for the problem described
in 7.3 given below:
Event Ei
E1 : 18
E2 : 19
E3 : 20
E4 : 21
E5 : 22
E6 : 23
7.5
Action Aj
A1 : 18
0
20
40
60
80
100
A1 : 19
50
0
20
40
60
80
A1 : 20
100
50
0
20
40
60
A1 : 21
150
100
50
0
20
40
A1 : 22
200
150
100
50
0
20
A1 : 23
250
200
150
100
50
0
Decision Models
Once the objective, alternative strategies and the decision making environments are known, the next step
which a decision maker faces is to select the decision model which can fit into his problem. Some of the
models are given below:
Deterministic Decision Model
Probabilistic Decision Model
Competitive Decision Model
Deterministic Decision Model

Deterministic model is related to deterministic situation. Deterministic decision payoffs are the simplest
possible payoffs. The objectives and strategies in this model have to be listed and then the payoff for each
strategy. If there are two objectives X1 and X2, the strategies selected are Y1 and Y2 and then payoffs are
shown below
158
Objectives
Strategies
Y1
Y2
X1
X2
Total Payoff
a11
a21
a12
a22
a1j
a2j
Where aij refers to payoffs of ith strategy towards jth Objective. In general, with m strategies
and n objectives, the decision payoff is as follows:
Objectives
Strategies
Y1
Y2
...
Ym
X1
X2
a11
a21
...
am1
a12
a22
...
...
am2
...
Xn
Total Payoff
a1n
a2n
...
amn
a1j
a2j
...
anj
Optimum strategy would be the one having the largest payoff. Some of the deterministic decision problems
can be solved by LPP.
Probabilistic Decision Model
Probability model is related to risk situation. Risk situation is one where there are many states and the
decision maker knows the probability of occurrence of each state. Decision payoffs are not fixed but
generally a random variable. Payoffs are determined partly by chance and partly by the strategy adopted.
Hence, a decision is made in favour of that strategy which has the maximum expected payoff.
Example 2. Suppose a businessman wants to stock commodity A or B. He can stock either but not both.
If he stocks A and if it is a success, he can make Rs. 200 but if it is a failure he will lose Rs. 500. If he stocks
B and if it a success then he can make 400 but if it is a failure he would lose Rs. 300. Which commodity A
or B should he stock? He has the following probability distribution.
Probability Of
With Commodity As stock
Success
0.8
Failure
0.2
Sol. Payoff matrix given below
With Commodity Bs stock

0.6
0.4
Success
Failure
A
+200
-500
B
+400
-300
Probability matrix is given in problem already. Now the expected payoff matrix is
Success
Failure
Total
Expected
Payoff
A
(0.8)(+200)
(0.2)(-500)
160-100 = +60
B
(0.6)(+400)
(0.4)(-300)
240-120 = +120
From the above table, it is clear that business should choose commodity B which has the
highest expected payoff = 120.
Decision Under Uncertainty

It is related to the situation of uncertainty. In an uncertainty situation, the probabilities of occurrence of the
different events are not known and the decision maker has no way of calculating the expected payoff for
his strategies. Consequently, there is no single best criterion for selecting a strategy to deal with such a
situation but there are different criteria available for selecting a strategy which are given below
159
a)
Lapalace decision rule
This rule is based on the assumption that the probabilities of different states of nature for a given strategy are
all equal. Considering these equal probabilities the expected payoffs will be calculated and then the strategy
with the highest expected payoff is selected. It is also known as the criterion of insufficient reason.
b)
Maximin or Minimax decision rule
This is the criterion of pessimism. Under this rule the decision maker is pessimistic. As such he selects that
strategy which gives largest of the minimum payoffs i.e. the maximum of the minimum. Gains or in case of a
loss (Regret) matrix, minimum of the maximum losses. When dealing with the costs, the maximum cost
associated with each alternative is considered and the alternative which minimises this maximum cost is
chosen. For example the min. profit associated with various strategies is as follows:
S1 : Rs. 520
S2 : Rs. 710
S3 : Rs. 250
S4 : Rs. 290
S5 : Rs. 310
S6 : Rs. 460
Since the maximum of these is Rs. 710, the strategy S2 is selected corresponding to the maximin principle.
(c) Maximax or Minimin decision rule
This is the criterion of optimism. Under this rule the decision maker is optimistic. As such he selects that
strategy which gives him the best possible payoff or best of bests. If it is a profit matrix, the decision maker
selects a strategy which gives him the highest of the maximum payoffs i.e. maximum of the maximum or in
case of loss (Regret) matrix, the minimum of minimum losses. For example maximum payoff associated with
the different strategies is given below:
S1 : Rs. 510
S2 : Rs. 740
S3 : Rs. 210
S4 : Rs. 290
S5 : Rs. 370
S6 : Rs. 980
The highest profit is Rs. 980, strategy S6 of ordering 40 copies of the magazines is the decision according to
the maximax principle.
(d) HURICZ decision rule
According to this rule, it is assumed that the decision maker is pessimist and (1) optimist. A linear
combination of the values pessimist and optimist are calculated for each action. The one with maximum
expected profit is chosen. Under these criteria the final choice is governed by the level of optimism/pessimism
of the decision maker i.e. the choice may differ for different decision maker according to their level of
optimism/pessimism.
160
(e) Savage decision rule

This rule is based on general insurance against risk. Under this rule, one selects the strategy which causes
minimum of the maximum possible losses. In this, payoff matrix is converted into a loss or regret matrix. In
each cell, we enter the difference between what the decision maker would have done if he had known which
outcome would occur and the choice represented by the cell. Once the regret matrix is formed, the minimax
criterion can be applied to it to select the best course of action.
Example 3. A newspaper boy purchases n number of copies at the rate of Rs. 8/- per copy and sells at the
rate of Rs. 10/- per copy. Any paper remaining unsold at the end of the day is disposed off as waste at Re.
1/- per copy. From past experience, it is known that the demand for such copies has always varied between
21-25. If all the copies are sold and there is a demand for more copies, the newspaper boy does not get
another chance to go and purchase additional copies to be sold. This problem is to decide the number of
copies that he should stock in the morning so as to maximize his profit.
Sol. Step 1: Set up the conditional Payoff Matrix as given below:
Possible Actions
State of Nature (Demand A1

of copies in the market)
(21)
A2
(22)
A3
(23)
A4
(24)
A5
(25)
D1 (21)
42
35
28
21
14
D2 (22)
42
44
37
30
23
D3 (23)
42
44
46
39
32
D4 (24)
42
44
46
48
41
D5 (25)
42
44
46
48
50
Table 1
The cell entries are calculated as follows:
e.g. for cell(2,3). 23 copies stocked, 22 sold (demand being 22)
22 copies sold @ 10/-
220
1 copy sold @ 1/-
Total Revenue
221
Cost of 23 copies @ Rs. 8/- =
184
Total conditional profit
37
Similarly calculations are done for all other cells.

Answer according to
a) Laplace Decision Rule:
Calculate average for each action for Table 1. This is known as expected value for each action.
A1
A2
(21) (22)
Average:
42
A3
(23)
42.2
A4
(24)
40.6
161
A5
(25)
37.2
32
(Expected Value)
Since expected value for Action A2 (22) is maximum, the decision is to stock 22 copies in the morning with
an average profit of Rs. 42.2/-.
b) Maxmin decision rule (Also known as Pessimistic Approach)
The decision maker identifies the minimum profit that he can get for each of his decision choice
(actions) and chooses the one which is maximum among these. e.g. with reference to Table 1
A1
A2
(22)
(21)
Min. profit:
42
A3
(23)
35
A4
(24)
28
A5
(25)
21
14
Choose maximum from among these.

42 being the maximum, the decision is to stock 21 copies in the morning with an average profit of Rs. 42/.
a) Maximax decision rule (Also known as Optimistic Approach)
The decision maker identifies the maximum obtainable profit for each of his decision choice (actions)
and chooses the one which is maximum among these.
A1
(21)
Max. profit:
A2
(22)
42
A3
(23)
44
46
A4
(24)
48
A5
(25)
50
Choose maximum from among these.

Maximum of the maximum being 50; the decision is A5 i.e. stock 25 in the morning for an average profit of
Rs. 50/- .
b) HURWICZ decision rule:
The calculations are as follows
Previously the maxmin and maxmax; pessimistic and optimistic calculations have been done.
Now suppose the level of pessimism is 0.3 and
Therefore the level of optimism is (1) = (1 0.3) = 0.7
Vp
V0
VH
A1 (21)
42
42
42
(42*0.3 + 42*0.7) = 42
A2 (22)
35
44
41.3
(35*0.3 + 44*0.7) = 41.3
A3 (23)
28
46
40.6
(28*0.3 + 46*0.7) = 40.6
A4 (24)
21
48
39.9
(21*0.3 + 48*0.7) = 39.9
A5 (25)
14
50
39.2
(14*0.3 + 50*0.7) = 39.2
Vp Value for action under pessimist criteria

Vo Value for action under optimist criteria
VH Value for action under HURWICZ criteria
According to the above calculations, the HURWICZ value VH (for =0.3 and (1) = 0.7) is 42 i.e.
maximum for A1.
Therefore according to this criterion the decision is to stock 21 copies in the morning with expected
profit of rs. 42/- .
162
c) Minimax Regret criteria

In this case first of all, using the conditional payoff matrix, a regret table is generated.
For each state of nature (rows in this case) all values of the row are subtracted from the maximum value of
the row. This is done for all rows. Accordingly the regret matrix is obtained as follows.
Possible Actions

(21)
A2
(22)
A3
(23)
A4
(24)
A5
(25)
D1 (21)
42
35
28
21
14
D2 (22)
42
44
37
30
23
D3 (23)
42
44
46
39
32
D4 (24)
42
44
46
48
41
D5 (25)
42
44
46
48
50
Possible Actions

(21)
A2
(22)
A3
(23)
A4
(24)
A5
(25)
D1 (21)
14
21
28
D2 (22)
14
21
D3 (23)
14
D4 (24)
D5 (25)
maximum regret for each

action
14
21
28
choose minimum regret
The minimum being 7, the decision is: A2 i.e. stock 22 copies in the morning with the average regret
value = 7.
This decision is same as was obtained under max min profit criteria. The decision resultant from maximum
profit and min max regret will always be same.
Decision under Risk
Risk situation is slightly better than uncertainty in the sense that in risk situation probability of occurrence of
States of Nature are known. The decision maker takes a well considered and calculated decision and risk.
Example 4: Let us consider the same example of newspaper boy with additional information that now
probabilities of States of Nature (demand) is known and stated. The resultant conditional pay off matrix is
as follows:
163
Decision Maker
Possible Actions:
(No. of copies he can stock)
States of Nature (Demand of

Copies in the Market)
P
A1
21
A2
22
A3
23
A4
24
A5
25
D1 (21)
0.1
42
35
28
21
14
D2(22)
0.15 42
44
37
30
23
D3(23)
0.45 42
44
46
39
32
D4(24)
0.2
42
44
46
48
41
D5 (25)
0.1
42
44
46
48
50
42
43.1 42.85 38.55 32.45
E (Ai)
Criteria used in decision making under uncertainty
(i) Criteria of Maximum likelihood.
(ii) Criteria of maximum expectation
(iii) Criteria of minimum expected regret / opportunity loss.

(i) Criteria of Maximum likelihood:
According to these criteria the decision maker chooses action for the event for which probability ofhappening
in maximum. For example: in the pay off matrix given Example 2; 45 is the maximum profitability for occurrence
of demand of 23 newspapers.
For demand of 23 news papers, his maximum profit is Rs 46/- against action A3 = Stock 23 papers.
Therefore he will stock 23 news papers in the morning.
(ii) Criteria of Maximum Expectation:
According to these criteria; the decision maker will choose that Action (Ai) for which it expected value is
maximum. For action Ai the expected value E (Ai) is calculated as follows:
A3 = 0.1 28 + 0.15 37 + 0.45 46 + 0.2 46 + 0.1 46
= 42.85
Similarly, the calculations for all Actions (columns) are completed. The calculated values are shown in last
row in table against.
Actions
A1
A2
A3
A4
A5
E (A1)
42
43.1
42.85
38.55
32.55
164
The maximum expected value among E (Ai) being 43% for action A2; the decision will be the stock 22
copies for an expected profit of 43.1 Rupees.
(iii) Criteria of Minimum Expected Regret
Consider the regret table generated in earlier case (Decision under uncertainty with minimum regret); along
with the probabilities considered in the previous example:
(Regret Table)
Possible Actions: No. of copies he can stock
P
A1
21
A2
22
A3
23
A4
24
A5
25
D1 (21)
0.1
14
21
28
D2(22)
0.15 2
14
21
D3(23)
0.45 4
14
D4(24)
0.2
D5 (25)
0.1
4.1
3.25
7.55
13.65
Expected
Action
Regret For Each
Minimum Regret
Therefore under these criteria, the choice is stock 22 copies with minimum expected regret of 3. The
decision under these criteria will always be the same as under criteria of maximum expectation, discussed
earlier.
Value of Perfect Information:
In the above discussion, under criteria of maximum expectation, the decision is
A2 to stock 22 copies with an expected value of 43.1 Rupees.
This means the decision maker can realize an average profit of Rs. 43.1 / day, if he uses the strategy over
large number of 6 days.
Now, suppose there is a market research expert who claims that he can exactly predict demand for the day.
So that the decision maker can make the best choice every day depending upon that days prediction. So, if
he is able to tell exactly each days demand; he is providing PERFECT INFORMATION to the decision
maker for the problem situation. Therefore on the basis of this PERFECT INFORMATION the decision
maker is able to take the best decision each day.
165

A1
A2
A3
A4
A5
D1 (21)
P
0.1
21
42
22
35
23
28
24
21
25
14
D2(22)
0.15 42
44
37
30
23
D3(23)
0.45 42
44
46
39
32
D4(24)
0.2
42
44
46
48
41
D5 (25)
0.1
42
44
46
48
50
42
43.1 42.85 38.55 32.45
E (Ai)
As per above table:

Expected value under Perfect Information:
EVPI = 42 0.1 + 44 0.15 + 46 0.45 + 48 0.2 + 50 0.1
= 46.1
Therefore, the value of perfect information
VPI = Expected value under perfect information Maximum expected value
(As calculated under criteria of maximum expectation)
= 46.1 43.1
= 3.0
Rs 3/- is the value of perfect information. It is not advisable to pay to a consultant or marketing researcher
or advisor a fee is more than its value; in this case it is Rs 3/-.
Thus, it can be said that value of information is the ADDITIONAL amount (over and above what could be
earned otherwise) that can be earned using the information.
Decision Tree
In some situations the decision problem can be represented graphically known as Decision Tree of Decision
Flow Diagram. This type of representation is more useful, where sequences of decisions are required to be
taken one after the other. This is termed as multistage decision problem. That means some times a decision
situation can be graphically represented as one stage, two stage or multistage problem; and accordingly
analyzed to arrive at a decision.
Example 5: A Company is currently working with a process, which, after paying for materials, labour and
so on brings a profit of Rs. 12,000. The company has the following alternatives:
(i)
The Company can conduct research R1 which is expected to cost Rs 10,000 and having 90%
probability of success. IF successful, the gross income will be Rs 26,000.
166
(ii)
The company can conduct research R2, expected to cost Rs 6,000 and having a probability of
60% success. It successful, the gross income will be Rs 24,000.
(iii)
The company can pay Rs 5,000 as royalty of a new process which will bring a gross income of
Rs 20,000.
Because of limited resources, only one of the two types of research can be carried out at a time. Draw the
decision tree and bring out the optimal strategy for the company.
Decision tree corresponding to the above example is shown below:
Continue with Present Process
12,000
Success
0.9
Conduct Research R1
16,000
A
Failure
II
0.1
-10,000
Success
III Conduct Research R2
18,000
0.6
B
Failure
-6,000
0.4
IV Pay Royalty
15,000
Analysis:
The situation has four alternatives I, II, III, IV as shown in the above figure. This is a one stage problem.
The decision point is shown as a box.
Small
show the possible outcome at those points;
A and B in this case:

The process of Analysis is shown in table below:
Alternative
Net Expected Outcome
I) Continue with Present Process
Given
II) Conduct Research R1
16,000 0.9 10,000 0.1 = 13,400
III) Conduct Research R2
18,000 0.6 6,000 0.4 = 8,400
IV) Pay Royalty for new process
20,000 5,000
167
= 12,000
= 15,000
Since the expected outcome 15,000 is highest among all the four alternatives, it is recommended that
alternative IV i.e. Pay royalty of 5,000/- for a new process to earn gross income of Rs. 20,000 be
selected.
Example 6. Aldico Ltd. has installed a machine costing Rs. 4 lacs and deciding an appropriate number of
a certain spare parts required. Cost of spare part each Rs. 4000 but are available only if they are ordered
now. If machine fails and no spare parts are available, the cost to the company of mending the plant would
be Rs. 18000. Estimated life of the plant is of 8 years and the probability distribution of failures during this
time is as follows
No. of failures during 8 yearly period
0
1
2
3
4
5
6
Probability
0.1
0.2
0.3
0.2
0.1
0.1
0
Determine the following

i)
The optimal number of units of spare parts on the basis of
(a)
Minimax principle
(b)
Minimin principle
(c)
Laplace principle
ii)
Expected number of failures in the 8 yearly periods.
iii)
Regret table
iv)
EVPI.
F - number of failures
S - number of spares
C total cost
Cost function can be stated as follows
C = 4000S
when F S
= 4000S + 18000 (F S) when F > S
Sol. Cost matrix shown below
No.
of Probability
Failures
F1 : 0
0.1
F2 : 1
0.2
F3 : 2
0.3
F4 : 3
0.2
F5 : 4
0.1
F6 : 5
0.1
Column Minima
Column Maxima
Simple Average Cost
Expected Cost
No. of Spares
S1 : 0
S2 : 1
0
4
18
4
36
22
54
40
72
58
90
76
0
4
90
76
45
34
41.4
29.2
S3 : 2
8
8
8
26
44
62
8
62
26
20.6
168
S4 : 3
12
12
12
12
30
48
12
48
21
17.4
S5 : 4
16
16
16
16
16
34
16
34
19
17.8
S6 : 5
20
20
20
20
20
20
20
20
20
20
i)
No. of units according to

a)
Minimax : The maximum values in each of the columns are shown by the column maxima
row. The minimum of these is 20. Hence the decision would be to buy 5 spare parts.
b)
Minimin : The minimum values in each of the columns are shown by the column minima row.
The minimum of these is 0. Hence the decision would be to buy nothing.
c)
Laplace Principle: According to this, decision is taken on the basis of the simple average
cost values. Since the simple average cost is minimum for S5. Hence the decision would be to
buy 4 spare parts.
ii)
Expected number of failures in the 8 year period.

6
E(F) =
pi Ei
i=1
Thus E(F) = 0.1*0 + 0.2*1 + 0.3*2 + 0.2*3 + 0.1*4 + 0.1*5 = 2.3
iii)
Regret Table is given below
No.
of Probability
Failures
F1 : 0
0.1
F2 : 1
0.2
F3 : 2
0.3
F4 : 3
0.2
F5 : 4
0.1
F6 : 5
0.1
Expected Regret
iv)
No. of Spares
S1 : 0
S2 : 1
0
4
14
0
28
14
42
28
56
42
70
56
32.2
20
S3 : 2
8
4
0
14
28
42
11.4
S4 : 3
12
8
4
0
14
28
8.2
S5 : 4
16
12
8
4
0
14
8.6
S6 : 5
20
16
12
8
4
0
10.8
Expected regret for S4 is the least. Hence the decision would be to buy 3 spare parts.
EVPI (Expected value of perfect Information)
EVPI = Expected cost with optimal policy Expected cost with perfect information
Expected cost with optimal policy is 17.4 thousand rupees (Given in problem)
Event
F1 : 0
F2 : 1
F3 : 2
F4 : 3
F5 : 4
F6 : 5
Expected Cost
Cost
0
4
8
12
16
20
Prob.
0.1
0.2
0.3
0.2
0.1
0.1
EVPI = 17.4 9.2 = 8.2 thousand rupees.
169
Prob. X Cost
0.0
0.8
2.4
2.4
1.6
2.0
9.2
7.6
Bayesian Decision Rule: Posterior Analysis
In this approach, the optimal strategy is chosen using the expected value criteria while the expected pay-offs
are calculated by using posterior probabilities. In this rule, prior information of the decision maker is revised
on the basis of some additional information i.e. the using this information, prior probabilities are converted
into the posterior probabilities.
Example 7: A construction company has recently acquired a piece of land and plans to construct a
shopping Mall. The company has to decide the size of the complex. There are three options:
a) A small sized complex with 40 condominiums and a multiplex
b) A medium sized complex with 60 condominiums and a multiplex
c) A large sized complex with 100 condominiums and a multiplex
The company feels that the overall demand for the condominiums built would be either high or low. The
returns from the project will depend on size of complex and what the level of demand eventually turns out to
be. The pay-offs expected under various event-action combinations, together with estimated probabilities
of the demand are given below
Table 1
Event
Probability
Actions Aj
Small Complex Medium Complex
Large Complex
High demand
0.4
1800
2200
4200
Low demand
0.6
1000
600
-1200
Company has the choice of engaging a marketing research company to conduct a survey so that it can take
a more informed decision. Outcome of research as follows:
A favourable report indicating high demand
An unfavourable report indicating low demand
The past record of the research firm has led to the following estimates of the relevant probabilities
Event
High demand
Low demand
Table 2
Marketing Research Report
Favourable
Unfavourable
0.9
0.1
0.2
0.8
The fee of marketing research Firm is Rs. 30000 for this study. How should the construction company
proceed?
Sol. The company has two options to decide on the size of the complex.
a) Prior analysis
b) Posterior analysis
Prior analysis: The pay-off table is reproduced from table 1 and is shown below
170
Event
Probability
High demand
0.4
Low demand
0.6
Expected Pay-off
Actions Aj
Small Complex
1800
1000
1320
Medium Complex
2200
600
1240
Large Complex
4200
-1200
960
Here the expected pay-off for Small Complex is maximum i.e. optimal course of action.
Event
E1 : High demand
E2 : Low demand
EPPI
Probability
0.4
0.6
Pay-off
4200
1000
Pay-off X Probability
1680
600
2280
EVPI EPPI EP 2280 1320 960
Posterior analysis: Let us consider how posterior probabilities may be calculated and used. It may be
noted that the probabilities given in past records are in fact conditional probabilities.
P(I1 / E1) - probability that research firm gives a favourable report given state of nature eventually turns out
to be higher demand.
P(I1 / E2) - probability that research firm gives a favourable report given state of nature eventually turns out
to be lower demand.
P(I2 / E1) - probability that research firm gives an unfavourable report given state of nature eventually turns
out to be higher demand.
P(I2 / E2) - probability that research firm gives an unfavourable report given state of nature eventually turns
out to be lower demand.
Total probability that the research report will be favourable
P( I1 ) P( E1 I1 ) P( E2 I1 )
= P (E1) X P (I1 / E1) + P (E2) X P (I1 / E2)
= 0.4*0.9 + 0.6*0.20
= 0.36 + 0.12
= 0.48
Total probability that the research report will be unfavourable
P( I1 ) P( E1 I 2 ) P( E2 I 2 )
= P (E1) X P (I2 / E1) + P (E2) X P (I2 / E2)
= 0.4*0.1 + 0.6*0.80
= 0.04 + 0.48
= 0.52
These posterior probabilities can be used to determine optimal course of action under
each of these situations.
171
a) When a favourable report is given

The posterior probabilities for the events E1 and E2 when a favourable report is given by marketing
research firm, are shown below
Event
Ei
E1
E2
Prior
P(Ei)
probability Conditional
Probability
P(I1 / Ei )
0.4
0.90
0.6
0.20
The posterior probabilities of the event, the
using these probabilities are shown below
Event Ei
Probability
pi
E1 : High demand
E2 : Low demand
Expected Pay-off
0.75
0.25
Actionj
Small
Complex
1800
1000
1600
Joint
Posterior
Probability
Probability
P(Ei I1)
P(Ei / I1)
0.36
0.36/0.48=0.75
0.12
0.12/0.48=0.25
pay-off matrix and the expected values
Medium
Complex
2200
600
1800
Large
Complex
4200
-1200
2850
Since the expected pay-off for the action Large Complex is highest. It is concluded that given a favourable
research report is obtained, the best course of action is large sized shopping complex.
b) When an unfavourable report is given
The posterior probabilities for the events E1 and E2 when an unfavourable report is given by marketing
research firm, are shown below
Event
Ei
Prior
P(Ei)
E1
E2
0.4
0.6
probability Conditional
Probability
P(I2 / Ei )
0.10
0.80
Joint
Probability
P(Ei I2)
0.04
0.48
Posterior
Probability
P(Ei / I2)
0.04/0.52=0.0769
0.48/0.52=0.9231
The posterior probabilities of the event, the pay-off matrix and the expected values using
these probabilities are shown below
Event Ei
E1 : High demand
E2 : Low demand
Expected Pay-off
Probability
pi
0.0769
0.9231
Actionj
Small
Complex
1800
1000
1061.52
Medium
Complex
2200
600
723.04
Large
Complex
4200
-1200
-784.74
Since the expected pay-off for the action Small Complex is highest. It is concluded that given a favourable
research report is obtained, the best course of action is small sized shopping complex.
Thus when a favourable market research report is obtained, the optimal course of action is Large Complex
and expected pay-off is of Rs. 2850.
When an unfavourable market research report is obtained, the optimal course of action is Small Complex
and expected pay-off is of Rs. 1061.52.
172
But these decisions are conditional in the sense that either of them can be taken only when the nature of
report is known. Expected pay-off value is shown below
Indicator
Probability
Favourable Report
Unfavourable Report
Expected Pay-off
0.48
0.52
Conditional Pay-off of Expected Value

the Best Action
2850
1368
1061.52
552
1920
Expected pay-off of the optimal decision is Rs. 1920 if it is based on the market research information as
against Rs. 1320 without such information.
Expected value of Sample Information (EVSI)
EVSI = Expected pay-off with sample information Expected pay-off without sample Information
= 1920 - 1320
= 600 thousand
The company can pay an amount of Rs. 600000 for research report. But firm has asked a fee of Rs.
300000. It is worth engaging it.
7.7
Limitations
1.
The prior knowledge of all the possible outcomes is necessary to be able to use the techniques /
methods used in decision theory.
2.
For construction of payoff table / regret table the monetary value of combination of actions &
possible states of nature; is to be known before hand.
3.
The decisions are probabilistic in nature, therefore the outcome of decisions not definite.
7.8
Summary
In this section you have learned how to develop payoff/ regret tables. Various techniques/methods for
taking decisions under uncertainty using the payoff/regret table have been discussed. These decisions are
under the assumption of single stage decision situations. The techniques available for decision under risk
(i.e. when the probabilities of the occurrence of events are known). The concept of value of information and
value of perfect information is also discussed .Finally the construction and analysis of decision tree is discussed.
Application of decision tree is comparatively more useful for multistage decision situations.
7.9
Key Words
Uncertainty: When the probabilities w. r. t. to happening of events (States of Nature) are not
known.
Risk : When the probabilities w. r. t. to happening of events (States of Nature) are known.
Pay of Matrix: A Matrix showing profit / loss for all combinations of Action and Possible events
(States of Nature).
173
Regret Matrix: A Matrix (table) matrix showing regret /opportunity loss for not having taken the
best decision.
Decision Tree: A graphical representation (in the form of branches) various decision points along
with the alternatives; and possible outcome with their probability events of happening. It also shows
in monitory term the profit / loss for each combination alternative event happening.

1.
Briefly explain the different decision rules adopted in decision making under condition of uncertainty.
2.
Discuss the main features of the decision theory. Also state its limitations.
3.
Write short note on the following

a)
Deterministic decision Model
b)
Probabilistic decision Model
c)
Decision Payoff
4.
Define EVPI. How it is calculated?
5.
Explain the posterior analysis of decision making.
6.
From the given matrix, elements of which indicate profits, obtain the decisions using the following
principles of decision making
a)
Maximax
b)
Maximin
c)
Laplace
a)
S1
S2
S3
S4
7.
a1
26
26
18
22
a2
22
22
22
22
a3
13
34
18
18
a4
22
30
18
18
a5
18
20
20
18
The Jeans manufacturer must decide whether to build a large factory or a small factory in a
particular location. The profit per pair of jeans manufactured is estimated as Rs. 10. A small factory
will have an amortised annual cost of rs. 200000 with a production capacity of 50000 jeans per
year. A large factory will have an amortised annual cost of Rs. 400000 with a production capacity
of 100000 jeans per year. Four levels of manufacturing demand are considered equally likely namely
10000, 20000, 50000 and 100000 pair of jeans per year.
a)
Set up a pay-off table for building a small factory and a large factory for manufacturing
designer jeans.
b)
Explain the meaning of EVPI.
c)
Which would you choose to build, a small or large factory? Why?
174
7.11 References
Quantitative Techniques in management by N.D.Vohra, Tata McGraw Hill.
Operations Research by S.D.Sharma, KEDAR NATH RAM NATH MEERUT DELHI.
Operations Research by Prem Kumar Gupta & D.S.Hira, S.Chand.
Operations Research by Hamdy A.Taha, Pearson Education.
Introduction to Operations Research by Billy E.Gillett, Tata McGraw Hill.
Operations Research Theory & Applications by J.K.Sharma, Macmillan India Ltd.
Operations Research by Richard Bronson, Govindasami Naadimuthu, Schaums Series.
175
Unit - 8: Operations Research

Unit Structure:
8.0
8.1
8.2
8.3
8.4
8.5
8.6
8.7
8.8
8.9
8.10
8.11
8.12
8.13
8.14
8.15
8.16
8.17
Objectives
Introduction
Definitions
Characteristics
Why Operations Research?
Applications of Operations Research Techniques
Operation Research Models
Principles of Modelling
Constructing a Model
Types of Mathematical Models
Management Applications of Operations Research
Scope of Operations Research
Development of Operations Research in India
Role of Computers in Operations Research
Summary
Key Words
References
8.0
Objectives
Meaning of Operations Research
Characteristics of Operations Research
Application areas of Operations Research

Operations Research Models, their advantage and disadvantages
How to construct an OR model
Various types of mathematical models
Operations Research scope
Operations research development in India
Impact of Computers in Operations Research
8.1
Introduction
Operation Research (OR) came into existence during the World-War II. In England, Military management
called upon a team of scientists to study the strategic and tactical problems at that time related to land and
air defence of the country. Their mission was to formulate specific proposals and plans for aiding the Military
commands to arrive at the decisions on optimal utilization of military resources and to implement the decisions
effectively also.
OR can be associated with AN ART OF WINNING THE WAR WITHOUT ACTUALLY FIGHTING
IT. OR teams were not actually engaged in military operations and in fighting the war. But they were
advisors and instrumental in winning the war to the extent that systematic approaches involved in OR
provides an intellectual support to strategic initiatives of military commands.
176
Following the end of war, the success of military teams attracted the attention of industrial managers who
were seeking solutions to their complex executive-type problems. The most common issue was - what
methods should be adopted for optimum deployment of limited resources so that total profit is maximised
and or total cost minimised ?
The impact of OR can be felt in many areas now a days. A large number of management consulting firms are
engaged in OR activities currently. OR activities include
Transportation system
Hospitals
Libraries
Financial Institutions etc. apart from military and business applications.
While making use of the OR techniques, a mathematical model of the problem is formulated. Actually this
model is a simplified representation of the problem. Then an optimal solution is found by utilising various
applicable mathematical operations. Since the model is an idealized form of exact representation of real
problem, most favourable solution thus obtained may not be the best. Although, at times, it may be possible
to develop, extremely accurate but complex mathematical models yet they may not be able to provide the
additional benefit. So from both the mathematical simplicity and cost minimising point of view, it seems good
to develop a simpler model. Thus the weaknesses in the initial solution are definitely used to suggest
improvements in the model. A new solution is obtained and the process is repeated until the further
improvements in the succeeding solutions become so small that it does not seem economical to make further
improvements.
If model is formulated carefully and tested, the resulting solution should be a good approximation of real
problem. Although we may not get the best answers but definitely we are able to find some what good
answers rather than not having any.
8.2
Definitions
There are several definitions of OR given by researchers
Year
Researcher
Name
Definition
1946
Morse and
Kimbal
OR is a scientific method of providing executive departments

with a quantitative basis for decisions regarding the operations
under their control.
1948
P.M.S.Blackett
OR is a scientific method of providing executive with an

analytical and objective basis for decisions.
1948
P.M.Morse
1957
Churchman,
Acoff, Arnoff
OR can be considered to be an attempt to study those

operations of modern society which involved organization of
men or of men and machines
OR is the application of scientific methods, techniques and
tools to problem involving the operations of systems so as to
provide those in control of the operations with optimum
solutions to the problem.
177
1958
T.L.Saaty
1968
Jagjit Singh
1968
Ackoff and
Sasieni
1975
Thiearf and
Klekamp
OR is the art of giving bad answers to problems to which

otherwise worse answers are given.
OR is a management activity pursued in two complementary
ways- one half by the free and bold exercise of commonsense
untrammelled by any routine and other half by the application
of a repertoire of well established procreated methods and
techniques.
OR is the application of scientific method by inter-disciplinary
teams to problems involving the controls of organized systems
so as to provide solutions which best serves the purpose of the
organization as a whole.
OR utilizes the planned approach and an inter-disciplinary team
in order to represent complex functional relationships as
mathematical models for purpose of providing a quantitative
basis for decision making and uncovering new problems for
quantitative analysis
Most of the definitions of Operations Research are not satisfactory due to three main reasons:
1. Operations Research is not a science like other scientific well defined phenomena i.e. well defined
physical, biological, social phenomena. Chemists know about atoms and molecules and have theory
about their interactions; biologists know about living organisms and have theory about their vital
processes, operations researchers dont claim to know or have theories about operations. It is
essentially a collection of mathematical techniques and tools which in conjunction with a system
approach are applied to solve practical decision problems of an applied nature.
2. Operations Research is inherently inter-disciplinary in nature and has applications in military, business,
medicine, engineering and so on. OR makes use of expertise and experience of people from
different disciplines for developing new methods and procedures. Thus, inter-disciplinary approach
is an important characteristic of OR which is not included in most of its definition.
3. Most of the definitions of OR have been offered at different times of development of OR and hence
these emphasize only one or the other aspect.
8.3
Characteristics
Various definitions of OR present different characteristics of OR. These are following:

i.
Application of scientific methods
ii. Use of inter-disciplinary teams

iii. System orientation
iv. Human factors
v. Quantitative solutions
vi. Use of computers
vii. Uncovering of new problems
viii. Improvement in the quality of decisions
178
8.4
Why Operations Research?
After studying in details about Operation Research, its characteristics, its need in industries can be studied.
Following factors contribute to the need and utility of Operations Research.
a) Complexity: In the present world, the number of factors influencing a decision has increased
drastically. Situations are becoming more complex and big because these factors are influencing
each other in a very complex manner. Hence, there is a great uncertainty about outcome of interaction
of factors like technological, environmental, competitive etc. Take an example of an industry process
i.e. production schedule which is taking into account:
a.
b.
c.
d.
e.
f.
Customer demand in critical situation.

Raw material requirement and availability.
Equipment capacity and probability of equipment failure.
Restriction on various manufacturing processes.
Number of operators/employees absent.
Power cut on that day.
Evidently, it is not easy to prepare a schedule which is both economical and realistic by considering
above factors without the help of some scientific technique. Thus, there is a need of mathematical
model which, in addition to optimization helps to analyse the complex situation.
b) Scattered responsibilities and authority: In the modern world when sizes of industries are very
big and responsibilities and authorities of decision makers are scattered throughout the organizations
and thus the organization if it is not conscious, may be following inconsistent goals. Mathematical
quantification or OR overcomes this difficulty to a great extent.
c) Uncertainty: There is a great uncertainty about economic and general environment. With economic
growth, uncertainty is also increasing. This makes decision making more difficult, costlier and time
consuming for the systems. OR, is, thus, quite essential from reliability point of view.
d) Knowledge explosion: The knowledge is increasing at various levels at a very fast rate. Majority
of the industries are not up-to-date with the latest prospects and are, therefore, at a disadvantage.
Various OR teams working for the industries gather latest information for analysis purposes which is
quite useful for the industries.
8.5
Applications of Operations Research Techniques
OR is being widely used in various types of industries, businesses, Governments, military, agricultural &
transportation sectors. A few industries / industrial sectors can be named such as: automotive, airlines,
railways, electronics, media, coal, petroleum, paper, chemical, software, etc. Few of the techniques which
are successfully applied are:
Inventory control: This model is used to sort out problems related to stocks by finding out Economic
Order Quantities, safety stocks in the inventories, minimum and maximum order quantities per lot
and availability of minimum and maximum stocks.
PERT & CPM network techniques: These programmes are used in various types of projects
like flyovers, railway tracks, dams, highways, roads for planning and scheduling of activities, also in
development and production of ships, aircrafts, helicopters, computers etc.
Queuing theory: This theory is used in solving problems which are related to scheduling of air
traffic, repair and maintenance of automotive industrial areas, bank counters, bus and railway bookings,
traffic congestion etc.
179
Linear Programming: This technique is widely used in solving the assignments of job allocation to
machines, materials, distribution, transportation etc.
Simulation: The theory being used in analyzing market situations where probability factors are
linked to sort out problems.
Dynamic Programming: It is applied in areas like selection of advertising media, cargo loading
and distribution, routing optimization etc.
Decision Theory: It has been used in analysing and finding methods for making decisions under
various situations of uncertainty and risk to meet the stated objectives.
8.6
Operations Research Models
A model in OR is used as a method of representation of an actual object or a situation. A model shows

relationships (directly or indirectly) and inter-relationships of action and reaction in terms of cause and
effect.
The main objective of a model is to provide a means for analyzing the behaviour of the system for the
purpose of improving its performance. In case, the system existence is futuristic then a model defines the
structure of this future system indicating the functional relationships of various elements in the system. A
model allows the examination of behaviour of a system without interfering with ongoing operations.
There are various characteristics which are the base for classification of models, these can be classified as:
Classification by structure:
i. Iconic models: It represents the system as it is by scaling up or down (i.e. by enlarging or
reducing the size). In other words, it is an image. For e.g. a toy tank is an iconic model of a real one.
Other examples can be photographs, drawings, maps etc. Similarly a globe; the diameter of earth
is scaled down but the shape of earth is approximately the same. Relative sizes of continents, seas,
rivers, mountains etc. are approximately correct.
ii. Analogue models: In this type of model, one set of properties is used to represent another set
of properties. After solving the problem, the solution is reinterpreted in terms of the original system.
For e.g. graphs are very simple analogues because distance is used to represent the properties such
as: time, number, percent, age, weight and many other properties.
iii. Symbolic (Mathematical) models: It employs a set of mathematical symbols (i.e. letters,
numbers etc.) to represent the decision variables of the system. These variables are related together
by means of a mathematical equation or a set of equations to describe the behaviour of the system.
Classification by Purpose :
i. Descriptive models: A descriptive model describes some aspects of a situation based on
observations, survey, questionnaire results or available data. For example: the result of an opinion
poll.
ii. Predictive models: In this type of model, what if type of questions can be answered, i.e. they
can make predictions regarding certain events. For e.g. based on survey results, television networks
using attempt to explain and predict the election results beforehand.
iii. Prescriptive models: Finally, when a predictive model has been repeatedly successful, it can
be used to prescribe a source of action. For example, linear programming is a prescriptive model
because it prescribes what the managers ought to do under the given circumstances.
180
Classification by nature of environment :

i. Deterministic models: These models assume conditions of complete certainty and perfect
knowledge. For e.g. transportation & assignment models are deterministic types of models.
ii. Probabilistic (or Stochastic) models: These models usually handle the situations in which the
consequences of managerial actions cannot be predicted with certainty. For e.g. insurance
companies insure against risk of fire, accidents, sickness etc.
Classification by Behaviour:
i. Static models: These models dont consider the impact of changes that takes place during the
planning horizon i.e. they are independent of time. Also, in a static model only one decision is
needed for the duration of a given period of time.
ii. Dynamic models: In these models, time is considered as one of the important variables and
admits the impact of changes generated by time. Also, in dynamic models, not only one but a series
of independent decisions is required during the planning horizon.
Classification by Method of solution:

i. Analytical models: These models have a specific mathematical structure and thus can be solved
by known analytical or mathematical techniques. For e.g. a general Linear Programming model and
specifically structured transportation models are analytical models.
ii. Simulation models: These models also have mathematical structures but they cannot be solved
by purely using the tools and techniques of mathematics.
Classification by Use of Digital Computers:

i. Analogue and mathematical models combined: Sometimes analogue models are also expressed
in terms of mathematical symbols. For e.g. simulation model is of analogue type but mathematical
formulae are also used in it.
ii. Function models: These models are grouped on the basis of the function being performed.
iii. Quantitative models: These models are used to measure the observations.
iv. Heuristic models: These models are mainly used to explore alternative strategies, or arrive at
some acceptable decision by using some rule or procedure. Such rules may not have sound
mathematical proofs to justify their use, but have been found to be useful in practice.
8.7
Principles of Modelling
Some principles are discussed which are useful in providing guidelines during when formulation of models
takes place. While model building, these principles can be kept in mind:
1. Better to make a simple model than a complicated one:
It is a common guiding principle for mathematicians who are in process of building a strongest
possible model that has wide applications.
2. Tendency to mould the problem to fit the technique:
For example an expert on linear programming techniques may tend to view every problem he
encounters requiring a linear programming solutions. In fact not all optimization problems involve
181
only linear functions. Also, not all OR problems involve optimization. Therefore, not all real-world
problems call for Operations Research. Of course everyone searches and depicts reality in his own
terms, so the field of OR is not unique in this regard.
3. Beware of over-selling a model:
This principle is of importance for the OR professional as most of the non-technical benefactors of
an operations researchers work are not likely to understand his methods. More technicality in the
method also increases the responsibilities and burden on the OR professionals to distinguish clearly
his role; as model manipulator and model interpreter. In those cases, where the assumptions can be
challenged, it would be dishonest to use the model.
4. A model should neither be pressed to do, nor criticized for failing to do that for which it was
never intended:
One example of this error can be taken as the use of forecasting model to predict so far into the
future that the data on which the forecasts are based have no relevance. Another example is the use
of certain network methods to describe the activities involved in a complex project. A model should
not be stretched beyond its capabilities.
5. A model should never be taken too literally:
For example, there is a requirement of sophisticated model for Indian economy and a lot of researcher
teams are working on the project and these teams are spending a great deal of time and expenses
to gather all information regarding the complicated relationships amongst factors. Under these
circumstances, it can be believed that the model is an excellent representation of the real system,
thus may have been in use under the misplaced impression that so much of time effort must have
resulted in a good model.
6. Models should be validated prior to implementation:
Taking example of monthly sales forecast of a particular commodity, it can be compared with
historic data for testing purpose. If there is no possibility of validation testing before it is implemented
then it can be implemented in phases for validation. We can understand by taking a new model for
inventory control, first the model can be implemented with few selected items while the older system
is retained for the majority of remaining items. If the model proves to be successful, more items can
be taken.
7. The deduction phase of modelling must be conducted rigorously:
When conclusions of the model come into the picture and they are inconsistent with reality, it means
there is some error in the assumptions which are being taken hence rigorous deduction of modelling
is necessary. One application of this principle can be discussed i.e. Programming of computers must
be done very carefully, Hidden bugs are very dangerous as they dont prevent the programmes
from running but simply produce results which are not consistent with the intention of the models.
8. Some of the primary benefits of modelling are associated with the process of developing
the model:
It is important to know that a model is never as useful to anyone else as it is to those who are
involved in building it up. The model itself is non-explanatory and does not provide itself full knowledge
of the system that the modeller should acquire in order to successfully build it. In some cases, the
sole benefit of the modelling occurs when it is being built and after completion of the model, it has no
further value.
182
9. A model cannot be any better than the information that goes into it:
A model can only manipulate data as in case of a computer also, which is provided to it. Models can
condense data or they can use the data to provide some more useful information or data but they
dont have the quality or capacity to generate it.
10. Models cannot replace decision makers:
It should not be understood that model can be used to provide optimal solutions free from human
subjectively. OR models can aid decision makers and therefore provide a lot of support to take
better decisions to be made. Undoubtedly, experience, intuition and judgement in decision making
are undiminished.
8.7.1 Characteristics of a Good Model
A good model should possess the following characteristics.
It should be easy and economical to construct.
The number of simplifying assumptions should be as low as possible.
It should be adaptable to parametric type of treatment.
The number of relevant variables should be as low as possible. This means the model should be
simple yet close to reality.
It should assimilate the system environmental changes without change in its framework.
8.7.2 Advantages of Model in OR

Following are the advantages of using model in Operation Research.
It helps in finding avenue for new research and improvements in a system.
It indicates the scope as well as limitations of a problem.
It provides a logical and systematic approach to the problem.
It represents the overall structure of the problem more comprehensible and helps in dealing the
problem entirely.
8.7.3 Limitation of Model

Following are the limitations of a model.
The validity of a model for a particular situation can be ascertained only by conducting experiments
on it
Models are only a representation of reality and should not be regarded as absolute in any case.
8.8
Constructing a Model
Construction of a model is multi-step process. In first step, formulation of the problem requires analysis of
the system under study. This analysis shows various phases of the system and the way it can be controlled.
With the formulation of the problem, the first stage in model construction is over. In next step, measure of
effectiveness is defined i.e. the next step is to construct a model in which effectiveness of the system is
expressed as a function of the variables defining the system. The general form of an OR model is:
E = f(x, y),
183
Where
E = Effectiveness of the system

x = Variable of the system that can be controlled
y = Variable of the system that cannot be controlled
Deriving of solution from such a model consists of determining those values of control variable x, for which
the measure of effectiveness is optimized. Optimization includes both maximization (in case of profits,
production capacity etc.) and minimization (in case of losses, cost of production etc.).
Various steps in the construction of a model are:
Selection of components of the system: All the components of the system which contribute
towards the effectiveness measure of the system should be listed.
Pertinence of components: Once a complete list of components is prepared, the next step is to
find whether or not to take each of these components into account. This is determined by finding the
effect of various alternatives courses of action on each of these components.
Combining the components: It may be convenient to group certain components of the system.
For e.g. material price, transportation charges of material and inspection cost of raw material can
be grouped together and called raw material acquisition cost. The next step is to find out, for the
components in modified list whether its value is fixed or variable.
Substituting symbols: Once each variable component in the modified list has been broken down
like this, symbols may be assigned to each of these sub-components.
8.9
Types of Mathematical Models
There are various types of mathematical models which are given below
Statistical Techniques:
Some of the techniques used in various functions, areas in this world come from Statistics &
Probability theories. A lot of factors in various areas cannot be predicted in absolute certainty hence
probability concept is associated with those factors to analyse the uncertainties in the situation and
data is used with some reasonable accuracies to reach some decision.
Queuing Models:
These models (may be used where the situations) involve the arrival of units to be serviced at one or
more stations. We can take example of car or any other vehicle reaching at loading station, patients
arriving any dental or other clinics, aeroplane reaching aerodromes, parcels reaching at P.A. desk,
customers entering a fair or a departmental store and so on. In case of single service station, there
will be a single queue and multiple queues can be used in multiple service stations (e.g. in a barber
shop). The same principle is applied in case a job is done through various processes on different
machines in workshop. In queuing model, there may be more demand of the material or less at
subsequent operation. The cost of waiting and idling of the material is different and needs to be
calculated and optimum queuing model to be selected to minimize the cost and improve the process.
Simulation Techniques:
This technique is a data generating technique which actually does not solve a problem but it generates
data or the information which is helpful in taking decisions. This method or technique is used in
cases when it is very much time consuming process to conduct an experiment or do actual real
184
study to find exact situation of a problem. Also, sometimes when there are large no. of variables are
available and there is lot of complexity between various variables. There is no possibility to develop
an analytical model representing the situation. For an example, there are two ways of developing
the project, the designer can make various aerodynamics equations of the plane and solve it and the
second way is to make a model of the aeroplane can be prepared and tested in a windmill.
Simulation is a very powerful method and one of the most widely used technique of operations
research. There are so many important decision-making problems which are too intricate to be
solved by mathematical analysis and the experimentation with the actual systems.
Competitive Models:
When two or more organizations or individuals are making decisions with conflicting objectives. In
these circumstances when one makes some decision, it will affect the decision made by another
individual(s). These models are applicable to situations like two players trying to win a game, planning
of war situations by two opponents, fight for a political election for a seat between two candidates,
companies trying to improvise share hold in the market etc.
Allocation Models:
These are used to solve the problems when there are so many jobs to be performed and there are
lot alternatives to complete the job at different stages with availability of resources are limited.
Simplest allocation (assignment) model is that in which no. of allocation of jobs is same as that ofno.
of resources (men). The assignment problem becomes more complex when some of the jobs require
more than one resource or if the resource can be used for more than one job.
Decision Theory:
Decisions are part of routine for mangers in organisations. The wrong decisions can create blunders
for the organisations and similarly right decisions may change / increase the profits marginally. There
may be two categories of decisions Strategic and Tactical decisions. Strategic decisions affect the
organisations in the long run and Tactical decisions affect the organisation in short term. New product
range development, site allocation, large scale automation etc. are the long term strategic decisions.
Before the decisions are taken by the manager, the data regarding the subject is collected, formulation
of data is done and all the possible alternatives are taken into account. Decision theory can make
use of probability factors of various outcomes. For example, a television manufacturing unit wants
to expand the production capacity, it can take various alternatives into consideration and select one
out of those. There may be following alternatives:
i. Increase capacity of the existing plant,
ii.Build a new production site,
iii. Hire production shop of other television company.
The main factor or uncertainty in the above situation is future demand of the television which can be
out of the following:
i.High demand
ii.Moderate demand
iii. Low demand
iv. No demand
Decision theory covers three categories of decision making:
185
i. Decision making under the conditions of certainty

ii. Decision making under conditions of risk
iii. Decision making under the conditions of uncertainty
Replacement Models:
There are lot of equipments, machinery, and parts being used by industries, railway and other
organisations like military etc. These machines and equipments get deteriorated with time and the
efficiency decrease with span of time. Running machines require more operational cost, running
vehicles need more maintenance costs. When the cost of maintenance and repair becomes more
expensive, it becomes necessity to replace the parts or the machines completely. There are no fixed
times when the replacements are needed. The replacement models help in making replacement
policies to calculate increased costs of repair, maintenance, scrap value and cost of equipment to
find the right intervals of replacements.
Similarly, there are few items like bulbs, tube-light etc. Which do not deteriorate but fail without
giving an alarm? The problem here is to find out which item to be replaced and at what intervals,
they should be replaced individually or in groups, the cost of handling, cost of replacement and cost
associated with failure of items to be calculated.
There is one more condition when the part needs replacement due to new inventions. In this situation,
the equipment is being replaced not because of not longer performance to the design standards
but due to because modern equipment performs to higher standards.
Sequencing Models:
These models are used in situations in which the effectiveness measure such as time, cost or distance
is a function of the order or sequence of performing a series of tasks. The selection of the appropriate
order in which waiting customers may be served is called sequencing. For e.g. x no. of machines are
used to do Xn type of operations and there are y no. of jobs to be allocated on machines. Now the
problem is to find out right sequence of operations so as to minimize total elapsed time on the
machines with condition that machines should be idle for minimum time.
Activities Scheduling by PERT & CPM:

There are cases when the no. of interrelated are huge in complex and big projects. In these projects,
management does not give the schedules of completion and in process just by intuition. Therefore,
some techniques are used by management teams for planning, scheduling and controlling the projects.
CPM and PERT are two techniques which are used to show diagrammatically representation of
various activities, their relations, dependencies etc. Time completions of these activities are shown
by arrows. There may be exact completion times of activities or probability of completion of those
activities. Depending on need and availability of data, different techniques are used.
Routing Models:
Routing problems are directly or indirectly related with sequencing. There are two important routing
problems:
The travelling salesman problem
The minimal path problems
186
In the travelling salesman problem, there may be a no. of cities where salesman will visit. The
distance between each cities is known, problem is to find out shortest distance of travel when the
person will start travelling from his home, go to different cities and come back to home city after
travelling different cities.
Inventory Models:
These models take data and experience of idle resources such as raw material, stock material,
machines, men, money etc. Models are concerned with two decisions:
How much quantity must be ordered &
When the order should be placed
So that the production should not stop, material shortages should not be there. Cost of ordering and
transportation should be minimized. Inventory should not go beyond minimum and maximum levels.
Cost of releasing order, follow up costs, storage space cost, capital and interest on capital cost
should be tried to minimize.
Reliability Theory:
Reliability is a factor which defines quality performance of the equipment, machine or a part. Reliability
is the capability of an equipment to perform without failure or malfunctioning for a specific time
period under given circumstances or environment. The equipment can be any simple one like fan,
electric heater, pulley, motor or any complex one like computer or an aeroplane.
Reliability of a machine or equipment depends upon the reliability of various parts of which it is
made. It is necessary that the equipment must perform the functions for which it is made. Otherwise,
the cost of failure can be very high and hazardous depending upon the use where it is being used.
For e.g. malfunctioning of software in space or in aeroplane during flight is unthinkable.
The reliability factor in quantitative term can be obtained for individual component by testing
procedures i.e. life testing. The reliability of the equipment can then be ascertained by using the
reliability of individual components and the design of equipment which indicates the interconnection
of components.
Advanced OR models:
Non-linear Programming:
If objective function and / or one or more of the constraints are non-linear in nature then nonlinear programming is used. Factors such as price discounts on bulk purchases, graduated
income tax etc. may cause non-linearity in the model. The non-linearity of the functions makes
the solution of the problem much more involved as compared to linear programming problems
and there is no single algorithm like simplex method which can be employed for situation of the
problem. Though, a number of algorithms have been developed, each is applicable to specific
type of non-linear programming problem only.
Integer Programming:
The method is used to solve the problems in which some or all variables must have nonnegative integer values. When all the variables are constrained to be integers, the problem is
called integer programming problem and in case only some of the variables are restricted to
have integer values, the problem is said to be a mixed integer programming problem. In some
187
situations, each variable can take on the values of either zero or one as in do or not do type
decisions. Such problems are referred to as zero-one programming problems.
Examples of integer programming are: number of trucks in a fleet, number of generators in a
power house. Approximate solutions can be obtained without using integer programming methods
but the approximation generally becomes poor as the values of the variables become smaller.
Dynamic Programming:
These models are used to make a series of interrelated decisions for multi-stage problems that
extend over a number of time periods. For example, a company may wish to make a series of
marketing decisions over time which will result in higher possible sales. Principle of the model
is that regardless of what the previous decisions are it tries to determine the optimum decisions
for the periods that still lie ahead. The dynamic programming approach divides the problem
into a no. of sub problems or stages. Though Dynamic programming models are used for the
problems in which time plays an important role, yet in many dynamic problems, time is not a
relevant variable. For instance, suppose a company has decide. capital C to be spent on its
advertisement of product and three advertising media can be used for the advertisement i.e.
radio, television and news paper. In each media, advertisement can appear a number of times
per week. Each appearance in the advertisement constituents some cost and the returns w.r.t
advertisement. Hence the problem is to determine, how many times, an advertisement should
be appeared in different alternate media per week so that the returns are at the maximum and
costs are within prescribed limits.
Goal Programming:
These models are similar to linear programming but are used in situations where there are
multiple goals and objectives. A company which is manufacturing lathes and milling machines,
there may be following objectives:
i. Maximize profits,
ii. Maximize number of lathes to be manufactured,
iii. Maximize number of milling machines to be manufactured.
It is obvious that all three goals cannot be added since units of measurement are decides targets
for each of the objective and ranks them in order of their importance. Receiving this information,
goal programming tries to minimize the deviations from the targets. It starts with the most important
goal and continues till the achievement of a less important goal.
Heuristic Programming:
This programming uses the rules of thumb to explore the most likely paths and to make educated
guesses to arrive at solution of a problem. This models seems to be quite promising for future
OR work. They bridge gap between strictly analytical formulations and the operating principles
which the managers are habitual of using. They go with step-by-step approach towards the
optimal solutions when a problem cannot be solved by mathematical programming form.
Quadratic Programming:
This refers to the problems in which objective function is quadratic in form (contains square
terms) while the constraints are linear. A number of efficient algorithms have been developed for
such problems.
188
Sensitivity Analysis:
After attaining optimal solution to a linear programming model, it may be desirable to study
how the current solution changes when the parameters of the problem get changed.
Parametric Programming:
The study of the effect of continuous changes in the values of the parameters on the optimal
solution to a linear programming problem.
Combined Methods:
Real systems may not involve only one of the models which are explained in above paragraphs. For
example, in production related problems, normally a combination of inventory, allocation and queuing
models are applied.
These combined models are solved once at a time and in some logical sequence. However, In
Operations Research, these models are combined and with the combination of those models, Master
model is built which considers many individual models.
At the last, it can be said that above mentioned / explained models solve a lot of problems in OR but
all the problems cannot be covered by these. It is also possible in near future, more new processes
will be revealed and subjected to more mathematical analysis.
Mathematical Techniques:
There is possibility that any mathematical technique can become a useful tool for operations analysis.
Most commonly used mathematical techniques may be used such as integral equations, differential
equations, vector and matrix theory etc.
8.10 Management Applications of Operations Research

Some of the areas of management of decision making where the tools and techniques of OR are applied can
be outlined as follows:
i.
Accounts Investments and Budgeting :

a. Cash flow analysis, long range capital requirements, dividend policies, investment portfolios
b. Credit policies and risks, delinquent account procedures
c. Claim and complaint procedures
ii. Purchasing, procurement and exploration :

a.
b.
c.
d.
Rules of buying, supplies; stable or varying prices

Knowing quantities, quality and times of
Replacement policies
Defining strategies for exploration of raw material sources
iii. Production Management :

a. Physical distribution
i. Location and sizes of warehouses, distribution centres and retail outlets
ii. Defining distribution policy
b. Facilities planning
i. Numbers and location of factories, warehouses, hospitals etc.
189
ii. Loading, unloading facilities for trucks, rails and determining transportation schedules
c. Manufacturing
i. Production scheduling and sequencing etc.
ii. Stabilizing production, employment training and layoffs
d. Maintenance and project scheduling
i. Maintenance policies and preventive maintenance.
ii. Project scheduling and allocation of resources
iv. Marketing:
a. Product selection, competency actions
b. Number of salesman, frequency of calling on account percent of time spent on prospects
c. Advertising media w.r.t. cost and time
v. Personel Management :
a. Selection of suitable personnel on minimum salaries
b. Mixes of age and skills
c. Recruitment policies and assignment of jobs
vi. Research and Development :
a.
b.
c.
d.
Determination of the areas of concentration of research and development

Project selection
Determination of time cost trade off and control of development projects
Reliability and alternative design
8.11 Scope of Operations Research

In the recent years of organized development, OR has entered successfully in many areas of research for
industry, government, education and military. Basic problem in developing countries is to remove poverty
as soon as possible. So there is great scope for administrator, economists, politicians and technicians working
in a team to solve problem by using OR technique. OR is also useful in the following fields:
1. In Defence:
OR has a lot of scope in defence operations. In modern phase, defence operations are divided into
Army, Airforce and Navy and further it is subdivided into operations, intelligence, administration,
training and the others. Therefore, to coordinate various departments and to arrive at optimum
strategy, OR is useful.
2. In Production Management:
Production manager can use OR techniques to:
Schedule the type of machine to be used for producing various items
Decide no. and size of items to be produced
Calculate optimum product mix
Select, locate and design the sites for production plant
3. In Marketing:
Marketing manager can decide:
190
Where to distribute product so that cost of transportation is optimized to minimum
Selection of right advertising media wrt time
4. In Industry:
There are a lot of processes through which material is passed in industries, starting from purchasing
of the material to various manufacturing processes and then despatch. The management is interested
in optimizing the costs of various processes to maximize profits. OR team studies all these processes,
compares the alternatives and fixes the right ones. OR has been used in fixing in various fields of
production, sales, purchases, inventory control, transportation, scheduling, sequencing, planning,
maintenance and repairs.
5. In Agriculture:
OR approach needed to be equally developed in agricultural sector nationally and internationally.
With increasing population all over and shortage of food due to unplanned behaviour, OR techniques
application was required. The problem of optimal distribution of water resources is faced by each
developing country and a good amount of scientific work can be done in this direction.
6. In Personnel Management:
Personnel manager can decide various activities using OR techniques:
Optimizing suitable no. of persons in minimum salary
Determining best age for retirement for the employees
To decide no. of persons to be appointed on full time basis for seasonal workload
7. In Finance:
In the modern world, it is necessary for every Government to utilise its resources for economic
development of the country:
To select best profitable plan of the company or the institution
To maximize per capita income
8.11.1 Role of Operations Research in Decision Making

Operations Research may be regarded as a tool which is used by the management to increase the efficiency
in the operations and effectiveness of their decisions. OR techniques can be used to find out location of the
factories, projects, and location of warehouses, transportation means and their tracks. The advantages of
usege of OR techniques in business and systems may be classified as:
1. Better control: It is very costly for big concerns management to provide routine decisions through
a lot of supervisors. OR approach helps executives to devote attention to more pressing situations.
OR approach helps in production scheduling and inventory management.
2. Better co-ordination: Sometimes OR has been useful in maintaining law and order situation out of
chaos. OR based planning becomes driving force in alarming situation of marketing decisions with
the limitation of manufacturing capabilities.
3. Better system: OR techniques are useful in analyzing and deciding warehouse establishment out
of various alternates.
191
4. Better decisions: OR models frequently yield actions that do help in intuitive decision making.
Sometimes human mind cannot incorporate all the important factor without the help of OR and
computer analysis.
8.12 Development of Operations Research in India

Existence of Operations Research in India is considered in year 1949 when a branch of OR was setup in
Hyderabad. At the same time, Prof. R.S.Verma from Delhi University setup an OR team in Defence Science
Laboratory. Prof. P.C.Mahalanobis established an OR team in Indian Statistical Institute, Kolkata to sort
out National survey and planning related problems. In year 1957, Operations Research society of India
was formed and the same society became a member of International Federations of Operations Research
Societies in year 1960. A lot of Journals are published by / in India like OPRESEARCH, Industrial
Engineering and Management, Material Management Journal of India, Defence Science Journal,
SCIMA.
University of Delhi was the first one to start OR education in India, which started M.Sc. in Operations
Research in 1963. Simultaneously, OR was added as a subject in MBA in Institute of Management at
Kolkata and Ahmadabad. In the recent years OR has become so much popular that it has been introduced
in almost all Institutes and Universities in various disciplines like Mathematics, Statistics, Commerce,
Economics, Management studies, Medical Science etc.
8.13 Role of Computers in Operations Research

Computers have played a very important role in the development and progress of Operations Research in
India. There are so much complex computations in Operations Research that without the help of computers
it can take weeks and months to yield the results manually. So the computers have become very essential
and integral part of OR. These days, OR and Computers methodology are developing simultaneously.
Computer software packages are being used for effective and rapid calculations which is necessary part of
OR approach to solve various complex problems. Some of the softwares are:
Quantitative System for Operations Management: This software is very user friendly and it,
itself has problem solving algorithms for operations management problems.
Quantitative System for Business Plus, version 3.0: This software can be used for solving
algorithms, doing basic statistics, non-linear programming and financial analysis for Operations
Research.
LINDO (Linear Interactive Discrete Optimization), developed by Linus Scharge Lindo.
Value STORM: MS Quantitative Modelling for Decision support: This is a special version
of Personal STORM version 3.0 developed to be used for Operations Research.
8.14 Summary
In this unit, you have studied what Operations Research is today. It also discusses what are the basic
reasons that most of the definitions of OR are not satisfactory .What are the characteristics of it and why it
was needed in industry? It also describes the various application area of Operations Research.It also discusses
about OR Model, taxonomy of OR model, its pros and cons and procedure of constructing an OR Model.
What are the characteristics a good model should possess.It illustrates the management application of OR,
its scope and its role in decision making. What are the recent developments of it in India and also role of
computers in operations research.
192
8.15 Key Words
Operations Research (OR): It is a scientific method of providing executive with an analytical and
objective basis for decisions.
Model A model is used as a method of representation of an actual object or a situation. A model

shows relationships (directly or indirectly) and inter-relationships of action and reaction in terms of
cause and effect.
Programme Evaluation & Review Technique (PERT) & Critical Path Method (CPM): are
two techniques of OR which are used to show diagrammatic representation of various activities,
their relations, dependencies etc.
Linear Programming (LP): LP technique in OR is widely used in solving the assignments of job
allocation to machines, materials, distribution, transportation etc.
LINDO (Linear Interactive Discrete Optimization): Computer software: Computer software

package which is used for effective and rapid calculations.

1.
What is Operations Research? Why it was needed in Industry?
2.
What are the characteristics of OR? How it plays an important role in an organisation for decision
making?
3.
Explain various applications of OR.
4.
Explain different types of models in OR, their characteristics, advantages and limitations.
5.
Is there any use of computers in OR techniques. Elaborate.
6.
Write a short note on

a) Scope of OR
b) Developments of OR in India
7.
Describe the various reasons that OR definitions are not satisfactory.
8.
What do you mean by OR Model? What are the principles of modelling? Describe the process of
constructing an OR Model.
8.17 References
193
Unit - 9 : Linear Programming and Transportation Models

Unit Structure:
9.0
9.1
9.2
9.3
9.4
9.5
9.6
9.7
9.8
9.9
9.10
9.11
9.12
9.13
9.14
9.15
9.16
9.17
Objective
Introduction
Conditions
General Formulation
Slack & Surplus Variables
Standard Form
Matrix Form
Few Important Definitions
Applications
Assumptions
Solution of LP
Transportation Problems
Tabular Representation
Solution of Transportation problem
Summary
Key Words
References
9.0
Objective
General Formulation of Linear Programming problem
Standard & Matrix form of LP problems
Application of Linear Programming
Limitations & advantages of Linear Programming Technique
Transportation Problems & its Matrix form
Tabular Representation of a Transportation Model
Solution of transportation problem
9.1
Introduction
In mathematical expression, a linear form can be defined as a1 x1+ a2 x2+.............+ an xn, where a1, a2, .........., an are
constants and x1, x2............., xn are the variables. The theory was observed by George Dantzig and his associates
in 1947 while working in U.S. Air Force department that a large no. of military programming and planning
problems could be formulated as maximising / minimising a linear form of profit / loss function whose
variables were restricted to values satisfying a system of linear constraints. The term programming may be
defined as a process of determining a particular programme or plan of action.
194
9.2
Conditions
Linear programming can be considered as one of the most important techniques developed in the field of
Operations Research (OR). Linear Programming can be used for optimization problems if the following
conditions are satisfied
1. A well defined objective function must be there (such as profit, cost or quantities produced) which
is to be either maximised or minimised and which can be considered as linear function of decision
variables.
2. There must be constraints to the extent of attainment of the objective and these constraints must be
capable of being expressed as linear equations or inequalities in terms of variables.
3. Alternate courses of action must be there. To understand through an example, an automotive part
can be produced on two different drilling machines, now problem is how much quantity must be
produced on different machines.
4. Decision variables should be interrelated to each other and must be non-negative. The non-negativity
condition of the variables shows that they are real life problems seeking solution.
5. The resources must be in limited supply.
9.3
General Formulation
The general formulation of the linear programming problem can be stated as

In order to find the values of n decision variables x1, x2........, xn, to maximize or minimize the objective function
Z = c1 x1+ c2 x2+.............+ cn xn
And also satisfy m constraints
a11 x1+ a12 x2+....+. a1jxj............+ a1n xn ( or ) b1
a21 x1+ a22 x2+....+. a2jxj............+ a2n xn ( or ) b2
ai1 x1+ ai2 x2+....+. aijxj............+ ain xn ( or ) bi
Am1 x1+ am2 x2+....+. amjxj............+ amn xn ( or ) bm

Where constraints may be in the form of any inequality ( or ) or even in the form of an equation (=) and
finally satisfy the non-negativity restrictions
x1 0, x2 0, ........., xj 0, .........., xn 0.
However, by convention, the values of right side parameters bi (i = 1, 2, 3 ............., m) are restricted to non
negative values only. It is important to note that any negative bi can be changed to a positive value on
multiplying both sides of the constant by -1. This will not only change the sign of all left side coefficients and
right side parameters but will also change the direction of the inequality sign.
195
9.4
Slack & Surplus Variables
Slack variables If a constraint has sign, then in order to make it an equality, we have to add something
positive to the left hand side.
The non negative variable which is added to the left hand side of the constraint to convert it into equation is
called the slack variable.
x1+ x2 2,
..........................(1)
2x1+ 4x2 5,
..........................(2)
x1, x2 0
We add the slack variables x3 0 & x4 0 on the left hand sides of above inequalities resp.,
Which gives
x1+ x2 + x3 = 2,
..........................(1)
2x1+ 4x2 + x4 = 5,
.......................... (2)
x1, x2 ., x3, x4 0
Surplus variables If a constraint has sign, then in order to make it equality, we have to
subtract something non-negative from its left hand side.
Thus the positive variable which is subtracted from the left hand side of the constraint to
convert it into equation is called the surplus variable.
For e.g., considering the following constraints
x1+ x2 2,
..........................(3)
2x1+ 4x2 5,
.......................... (4)
x1, x2 0
We subtract the surplus variables x3 0 & x4 0 on the left hand sides of above inequalities
resp., Which gives
x1+ x2 - x3 = 2,
..........................(1)
2x1+ 4x2 - x4 = 5 ..................................(2)

...(2)
x1, x2, x3, x4 0
9.5
Standard Form
The standard form of linear programming problem is used to develop the procedure for solving general
linear programming problem. The characteristics of the standard form are explained in the following steps
Step I All the constraints should be converted to equations except for the non-negativity restrictions which
remain as inequalities ( 0). Constraints of the inequality type can be changed to equations by adding or
subtracting the left side of each such constraint by non-negative variables. These new variables are called
196
slack variables and are added if the constraints are ( ) or subtracted if the constraints are ( ). Since in the
case of constraint, the subtracted variable represents the surplus of the left side over the right side, it is
common to refer to it as surplus variable. For convenience, however, the name slack variable will also be
used to represent this type of variable. In this respect, a surplus is regarded as a negative slack.
For example, consider the constraints
3x1- 4x2 7, x1+ 2x2 3
These constraints can be changed to equations by introducing slack variables x3
Thus we get
3x1- 4x2 - x3= 7,
&
x4 resp.
x1+ 2x2 + x4 = 3 & x3, x4 0
Step II The right side element of each constraint should be made non-negative (if not). The
right side can always be made positive on multiplying both sides of the resulting equation by
(-1) whenever it is necessary. For example, consider the constraint as
4x1-4x2 -4 which can be written in the form of equation 3x1-4x2 - x3= -4 by introducing the
surplus variable x3 0
Again multiplying both sides by (-1), we get -3x1+4x2 + x3= 4 which is the constraint equation
in standard form.
Step III All variables must have non-integer values
A variable which is unrestricted in sign (that is, positive, negative or zero) is equivalent to the
difference between two non-negative variables. Thus if x is unconstrained in sign, it can be
replaced by (x, -x) where x and x are both non negative, that is x, x 0
The objective function should be of maximization form.
The minimization of a function f(x) is equivalent to the maximization of the negative
expression of this function, f(x), that is,
Min. f(x) = Max. [-f(x)]
For example, the linear objective function
Min. Z = c1 x1+ c2 x2+.............+ cn xn
Is equivalent to Max (-z) i.e. Max. z = -[c1 x1+ c2 x2+.............+ cn xn] with z = -z
Consequently, in any L.P. problem, the objective function can be put in the maximization
form.
By applying above mentioned steps systematically to general form of L.P. problem with all
constraints, the following standard form is obtained. Of course, no difficulty will arrive to
convert the general LPP with mixed constraints ( = )
Max. z = c1 x1+ c2 x2+.............+ cn xn+ 0 xn+1...... + 0 xn+m

197

a11 x1+ a12 x2+....+. a1nxn+ xn+1
b1
a21 x1+ a22 x2+....+ a2nxn+ xn+2
b2
am1 x1+ am2 x2+....+ amnxn............ + xn+m
bm
where. x1 0, x2 0,.... xn 0, xn+1 0.... xn+m 0
i.
This should be noted that the coefficient of slack variables xn+1 = 0, xn+2 = 0.... xn+m = 0 (are
assumed zero) in objective function so that the conversion of constraints to a system of simultaneous
linear equations does not change the function to be optimized.
ii. Since in the case of ( 0) constraints, the subtracted variable represents the surplus variable.
However, the name slack variable may also represent this type. In this respect, a surplus is regarded
as a negative slack.
9.6
Matrix Form
The linear programming problem in standard form [(3.11), (3.12), (3.13)] can be expressed in matrix form
as follows
Maximize z = CXT
(objective function)
subject to, AX = b, b 0
(constraint equation)
X0
(non-negative restriction)
X = (x1, x2,........ xn, xn+1...... ,xn+m),

C = (c1, c2, .....cn, 0, 0......, 0 ) and
b = (b1, b2, .....bm )
a11
a12
.......
a1n
......
a21
a22
.......
a2n
......
......
A=
am
am2
...... amn
....... 1
Similarly, treatment can be adopted in case of mixed constraints ( = ), following points will make this
clear.
The vector x is assume to include all decision variables, (i.e. original, slack and surplus). For convenience,
x is used to represent all types of variables. The vector C gives the corresponding coefficients in the objective
function. For example, if the variable is slack, its corresponding coefficient will be zero.
The Outlines of formulation of LP problems are explained with the help of these examples
198
Example 1. A company is making a chart to decide the minimum amount of constituents like proteins,
vitamins, carbohydrates, fats etc. which a man needs on daily basis to fulfil his requirement for medical
awareness. The choice is to be made from different type of foods (4 types). The yields per unit for different
types of foods are explained below in the chart. Formulation of the linear programming model is required for
this problem.
Type of food
Proteins
Yield per unit

Fats
Vitamins
6
8
16
12
1600
6
6
21
15
400
12
8
14
8
1400
A
B
C
D
Minimum reqd.
Cost per unit

(Rs.)
135
120
170
130
Sol. Formulation of L.P. Model

Let x1, x2, x3, x4 be the no. of units of food of type 1, 2, 3 & 4 resp.
Objective is to minimize the cost i.e.
Minimize Z = Rs. (135x1+ 120x2+ 170x3+ 130x4)
Constraints are on the fulfilment of daily requirements of the various constituents
6x1 + 8x2+ 16x3 + 12x4 1600
i.e. Proteins requirement
6x1 + 6x2 + 21x3 + 15x4 400
Fats
12x1 + 8x2 + 14x3 + 8x4 1400
Vitamins
Where each x1, x2, x3, x4 0

Example 2. Two types of products are manufactured by an organisation P1 & P2. They are sold at a profit
of Rs. 4 & 6 resp. Products are manufactured on machines M1 & M2 Product P1 requires to be
processed for 2 min. on machine M1 & 4 min. on machine M2. Product P2 is processed for 2 min. on
machine M1 & 2 min. on machine M2.
Machines M1 can be used for 800 min. a day & machine M2 is available for 600 min. a day. Formulation
of the linear programming model is required for this problem.
Sol. Let x1 be the no. of products of type P1 and, x2 the no. of products of type P2. After studying the
problem, the given information can be systematically represented in the table form as below
MACHINE
M1
M2
Profit per unit
(Rs.)
Type of products (minutes)

Type P1
Type P2
2
4
Rs. 2
2
2
Rs. 3
199
Available time
(min.)
800
600
Since profit on selling P1 is Rs.2 per unit, 2x1 will be profit on selling products of type A
Similarly profit on selling P2 is Rs.3 per unit, 3x2 will be profit on selling products of type B
Total profit is given by P = 2x1 + 3x2
Since P1 is manufactured on M1 for 2 min. and P2 on the same machine for 2 min., total no. of minutes on
machine M1 is given by
2x1 + 2x2
Similarly P1 is manufactured on M2 for 4 min. and P2 on machine for 2 min., total no. of minutes on machine
M2 is given by
4x1 + 2x2
But machine M1 is not available for more than 800 minutes, therefore
2x1 + 2x2 800
But machine M2 is not available for more than 600 minutes, therefore
4x1 + 2x2 600
Also, it is not possible to produce negative results for quantities, hence
x1 e 0 & x2 0
Hence allocation of the problem can be represented in the form
Find x1 & x2 such that the profit P = 2x1 + 3x2 is maximum,

subject to the conditions
2x1 + 2x2 800, 4x1 + 2x2 600, x1 0 & x2 0
Example 3. A toy company is supplying two types of aeroplanes a basic model of plane P1 & second
deluxe model of plane P2. Each plane of deluxe model takes twice as long as to produce as one of basic
type. The company would have time to make a maximum of 2000 per day. Supply of raw material is
sufficient to produce 1500 planes per day (both P1 & P2 combined).The deluxe model requires a high quality
wheel set of which there are 600 sets available per day only. They are sold at a profit of Rs. 3 & 6 per piece
resp. Then how many of planes of each type should be manufactured to maximize the total profit.
Formulation of the linear programming model is required for this problem.
Sol. Let x1 be the no. of planes of type P1 and, x2 the no. of planes of type P2. Let plane P1 requires t hrs, then
P2 requires 2t hrs. After studying the problem, the given information can be systematically represented in the
table form as below
Since profit on selling P1 is Rs. 3 per unit, 3x1 will be profit on selling planes of type P1
Similarly profit on selling P2 is Rs. 5 per unit, 5x2 will be profit on selling products of type P2
Total profit is given by P = 3x1 + 5x2
Since P1 is manufactured for t hrs. and P2 2t hrs. min., total time for manufacturing is given by
tx1 + 2tx2
But maximum time production is 2000 hrs, therefore
200
tx1 + 2tx2 2000t

x1 + 2x2 2000
But wheels sets are not available for more than 600 for deluxe model, therefore
x2 600
Also, it is not possible to produce more than 1500 quantities, hence
x1 + x2 1500
Also, it is not possible to produce negative results for quantities, hence
x1 0 & x2 0
HENCE allocation of the problem can be represented in the form
Find x1 & x2 such that the profit P = 3x1 + 5x2 is maximum,

x1 + 2x2 2000
x1 + x2 1500
x2 600
x1 0 & x2 0
Example 4. A farmer has 100 acre of farm for harvesting. potatoes, carrots & beans produced by him are
sold by him at the rates Rs. 2.00, Rs. 1.50 and Rs. 4.00 resp. The average productivity per acre of potatoes
is 4 tonnes, carrots are 6 tonnes and 2 tonnes of beans is produced per acre. Fertilizer is available at rate
of Rs. 1.0 per kg. and amount of fertilizer required for potatoes and carrots is 200 kg per acre and 100 kg
for beans. Labour required for sowing, cultivating and harvesting per acre is 5 man days for potatoes and
carrots, and 6 man-days for beans. A total of 400 man-days of labour are available at Rs. 60 per man day.
Formulation of the linear programming to maximize farmers total profit.
Sol. How much area should be given to each type of crop for cultivation so that profit from the production
can be maximized, it is to be decided.
Let x1, x2, & x3 be the acres of land for different crops (potatoes, carrots, and beans resp.). Therefore the
crops production will be 4000x1 of potatoes, 6000x2 of carrots & 2000x3 of beans
Total sales = [2 x 4000x1 + 1.5 x 6000x2 + 4 x 2000x3] = 8000x1+ 9000x2+ 8000x3
Total expenses on fertilizers = [1 x {200x(x1 + x2) + 100x3}] = 200x1+ 200x2+ 100x3
Labour expenses = [60 x {5x(x1 + x2) + 6x3}] = 300x1+ 300x2+ 360x3
Net profit (P) = Total sales Total expenses
= 8000x1+ 9000x2+ 8000x3 - 500x1- 500x2 -460x3 = 7500x1+ 8500x2+ 7540x3
Since total area is restricted to 100 acres of land, x1+ x2+ x3 100
201
Also, the total man-days labour is restricted to 400 man days, therefore
5x1+ 5x2+ 6x3 400
Hence the farmers allocation problem can be finally put in the form
Find the values of x1, x2, x3 so as to maximize

P = 7500x1+ 8500x2+ 7540x3,
x1+ x2+ x3 100
5x1+ 5x2+ 6x3 400
Where each x1, x2, x3
9.7
Few Important Definitions
There are few definitions which may help in understanding of forthcoming topics
i.
Solution to LPP: Any set of the type X = (x1, x2,........ xn, xn+1...... ,xn+m) of variables is called a solution to
linear programming problem if it satisfy the following constraints
a11 x1+ a12 x2+....+. a1nxn+ xn+1
a21 x1+ a22 x2+....+ a2nxn+ xn+2
am1 x1+ am2 x2+....+ amnxn............ + xn+m
ii.
b1
=
=
=
b2
bm
Feasible solution: Any set of the type X = (x1, x2,........ xn, xn+1...... ,xn+m) of variables is called a solution
to linear programming problem if it satisfy the above (in i) constraints and non negativity restrictions
also mentioned below
where x1 0, x2 0,.... xn 0, xn+1 0.... xn+m 0
iii.
Basic Solutions: A basic solution to the set of constraints (in i) is a solution obtained by setting any
n variables (among m+n variables0 equal to zero and solving for remaining m variables, provided
the determinant of the coefficients of these m variables is non-zero. Such m variables (of course,
some of them can be zero) are called basic variables and remaining n non zero-valued variables are
called non-basic variables.
The no. of basic solutions obtained will be the most m+n Cm = (m+n)! / n! m!, which is the number of
combinations of n+m things taken m at a time.
iv.
Basic Feasible Solution: A basic feasible solution is a basic solution which also satisfies the nonnegativity restrictions x1 0, x2 0, ...............xn 0, xn 1 0,......xn m 0 i.e. all basic variables
are non-negative. These are of two types
a. Non-degenerate BFS: A non-degenerate basic feasible solution is the basic feasible solution
which has exactly m positive xi (i = 1,2,3........,m). In other words, all m basic variables are positive
and the remaining n variables will be all zero.
202
b. Degenerate BFS: A basic feasible solution is called degenerate, if one or more basic variables
are zero-valued.
v. Optimum Basic Feasible Solution : A basic feasible solution is said to be optimum if it also
optimizes (maximizes or minimizes) the objective function
vi.
Unbounded solution: If the value of the objective function z can be increased or decreased indefinitely,
such solutions are called unbounded solutions.
9.8
Applications
In the present world, most of the events are non-linear type, yet some of the events occurring in routine life
are linear type or can be so approximated. Therefore, linear programming and its application must be
understood by the managers in solving the problems.
LP programming is widely used in various fields such as military, businesses, industries, marketing, distribution
and advertising problems to provide solutions. Reasons for its wide usage are
1. In L.P. models, a lot of varying data can be handled very easily.
2. A lot of powerful and efficient techniques are available to solve problems.
3. From various fields, problems can be easily represented through linear programming problems.
L.P. needs a lot of data to be handled which is complex and elaborative, in present scenario of computer
world, this disadvantage has been overcome completely as these computers can handle even large no. of
variables in L.P. problems in a short duration which can take days, months and even years if it is done
manually.
Application of Linear Programming in different areas are given below
Environment Protection
A lot of pollution in the form of carbon particles in air, waste material in water resources is taking place. L.P.
helps in analysis of different alternatives for efficient waste disposal, recycling of the material and energy
saving and utilising policies.
Transportation Problems
We can understand by taking an example, x no. of products / equipments / parts to be transported to y no.
of destinations. Cost of transportation to each destination, demand of destinations and storage costs of each
destination are known. The problem is to design the most appropriate transportation plan with schedule and
quantities which minimizes the total transportation cost.
Assignment Problems
These are used to solve the problems when there are so many jobs to be performed and there are lot of
alternatives to complete the job at different stages (with availability of resources are limited). Assignment
model is that in which no. of allocation of jobs is same as that of no. of resources (men). The assignment
problem becomes more complex when some of the jobs require more than one resource or if the resource
can be used for more than one job. Time required to complete the job at each source is known. Now,
optimum solution is to be found out which gives total minimum time to perform all the jobs.
203
Diet problems
It is another important category where linear programming is applied. Nutrient contents such as fats,
carbohydrates, vitamins, proteins, starch etc. is known in various food products. Minimum daily requirement
of each nutrient in the diet as well as the cost of each type of food stuff is given and the problem is to
determine the minimum cost diet which will satisfy the minimum nutrients requirement on daily basis.
Manpower scheduling problems
These types of problems are occurring in a lot of areas like big industries, hotels, hospitals, ports, etc. To
utilise the manpower resources in the most optimum manner, LP. Techniques are used which helps in reducing
the overtime costs (& other associated costs like transportation, wages etc. by allocating right no. of manpower
in different shifts (day, night etc.).
Agricultural problems
It is a very big area where there are lot of resources are used like seeds, machines for harvesting, fertilisers,
labour, land and capital to invest. The concern is to maximise the revenues and minimising costs.
9.9
Assumptions
Following are the assumptions for linear programming problem.

Additivity: To understand this,
if a machine takes t1 time to produce one unit of item P1
and t2 time to produce one unit of item P2,
then time required to produce one item each of P1 & P2 is t1+t2.
Proportionality: It is the primary requirement of linear programming problem that the objective function
and every constraint function must be linear. In simple words, if 1 kg of any item costs Rs. 5, then cost of 20
kg of the item will be 100 Rs. If the capacity of any steel plant is 1000 MT per day, it can produce
25000MT (25 x 1000) in one month (for 25 days working).
Multiplicativity: For producing one item on machine, it takes 2 hrs.,10 items of similar type will be
produced in 20 hrs.
Divisibility: It means that the fractional levels of variables must be permissible besides integral values. This
also means the variables are assumed to be continuous.
Deterministic: All the parameters in the linear programming models are assumed to be known exactly.
While in actual practice, production may depend upon chance also.
9.10 Solution of LP
Once a problem is formulated as mathematical model the next step is to solve the problem to get the optimal
solution. A linear programming problem with two variables can be solved by graphical method. The procedure
of graphical method can be understood as follows
i.
Each inequality constraint should be considered as equation.
ii.
Each equation must be plotted on the graph and each equation will be geometrically represented by
a straight line.
204
iii.
Feasible region will be shaded. Every point on the line will satisfy the equation of the line. If the
inequality constraint corresponding to the line is than the region below the line lying in the first
quadrant (due to non-negativity of variables) is shaded. For the inequality constraint with
sign, the region above the line in the first quadrant is shaded. The points lying in the common region
will satisfy all the constraints simultaneously. The common region thus obtained is called the feasible
region.
iv.
The convenient value of z will be chosen (say = 0) and the objective function line can be plotted.
v.
The objective function line will be pulled until the extreme points of the feasible region. In the
maximization case, this line will stop farthest from the origin and passing through at least one corner
of the feasible region. In the minimization case, this line will stop nearest to the origin and passing
through at least one corner of the feasible region.
vi.
The coordinates of the extreme points can be read selected in step (v) and find the maximum or
minimum value of z.
Graphical Procedure can be understood by the following examples

Example 1 Find a geometrical interpretation and solutions as well for the following LP problem.
Max. z = 3a1 + 5a2 subject to restrictions
a1 +2a2 2000,
a1 + a2 1500,
a2 600 and
a1, a2 0
Graphical Solution
Step i We will consider two mutually perpendicular lines ox1 and ox2 as axes of coordinates. Obviously, any
point (a1 , a2) in the positive quadrant will certainly satisfy non-negativity restrictions a1, a2 0. To plot the
line a1 + 2a2 = 2000, put a2 = 0, find a1 = 2000 from this equation.
We will mark a point L such that OL = 2000 by assuming a suitable scale, say 500 units = 2cm. Similarly,
when we put a1 = 0 to find a2 =1000 and mark another point M such that OM = 1000. Now we will join the
points L & M. The line will represent the equation a1 + 2a2 = 2000 as shown in the figure
a1 + 2a2 = 2000
M
Series 1
1600
1400
1200
1000
800
Series 1
600
400
200
0
0
500
1000
Fig. 1.1
1500
205
Point P lying on or below the line a1 + 2a2 = 2000 will satisfy the inequality a1 + 2a2 2000 (if we take a
point (500, 500), i.e. a1 = 500, a2 = 500, then putting the values in the equation, we have 500 + 2 x 500
2000, which is true)
Similar procedure is now adopted to plot the other two lines
Series 1
2000
1500
1000
Series 1
500
0
0
500
1000
1500
Fig. 1.2
and a2 = 600
Any point on or below the lines a1 + a2 = 1500 and a2 = 600 will satisfy the given.
Inequalities
a1 + a2 1500
and a2 600 respectively..
Series 1
1000
800
600
Series 1
400
200
0
0
500
1000
1500
2000
Fig. 1.3
Step II We will find the feasible region or solution space by combining the figures 1.1, 1.2 and 1.3. A
common shaded area OABCD can be obtained which is a set of points satisfying the inequality constraints
a1 + 2a2 2000 and a1 + a2 1500, a2 600 and non-negativity restrictions as a1, a2 0
Hence any point in the shaded area (including the boundary) is feasible solution to given LPP.
206
1600
1400
1200
a1 + a2 = 1500
1000
800
a2 = 600
C
600
400
a1 + 2a2 = 2000
200
0
500
1000
1500
Step III Co-ordinates of the corner points of feasible region O, A, B, C, D can be found out.
Step IV The corner points of optimal solution can be located either by calculating the value of z for each
corner point O, A, B, C and D (or by adopting the following procedure)
Here, The problem is to find out the point or points in the feasible region which maximize the objective
function. For some fixed value of z, z= 3a1 + 5a2 is a straight line and any point on it gives the same value of
z. Also, it should be noted that the lines corresponding to different values of z are parallel because the
gradient (-3/5) of the line z = 3a1 + 5a2 remains the same throughout. For z = 0, i.e. 0 = 3a1 + 5a2 , means
a line which passes through the origin. To draw the line 3a1 + 5a2 = 0 determine the ratio a1 / a2 = -5 / 3 =
-500 / 300
We can make a point E moving 500 units distance from the origin on the negative side of a1 axis. We may
find the points F such that EF = 300 units in the positive direction of a2 axis. Joining the point F & O, we can
draw the line 3a1 + 5a2 = 0 . We draw the lines parallel to this line until at least a line is found which is farthest
from the origin but passes through at least one corner of the feasible region at which the maximum value of
z is attained. It is also possible that such a line may coincide with one of the edge of feasible region. In that
case, every point on the edge gives the maximum value of z.
In this example, maximum value of z is attained at the corner point B (1000, 500) which is the point of
intersection of lines a1 + 3a2 = 2000 and a1+ a2 = 1500. Hence the required solution is a1 = 1000, a2 = 500
and max. value of z = Rs. 5500
Simplex Method is an iterative procedure for solving linear programming problem. It consists of the
following steps
a)
Having a trial basic feasible solution to constraint-equations
b)
Testing whether it is an optimal solution
c)
Improving the first trial solution by a set of rules and repeating the process till an optimal
solution is obtained.
The procedure requires at most m(= no. of equations) non-zero variables in the solution. One of the nonbasic variables at one iteration becomes basic at the following iteration is called an entering variable and one
207
of the basic variables at one iteration becomes non-basic at the following iteration and is called a departing
variable.
For convenience, construct the LP problem in standard form
Max. z = c1x1+ c2x2+ c3x3+ -----------------+ cnxn+ 0xn+1+0xn+2+-------------+ 0xn+m
S.T.
a11x1 + a12x2 + ---------------------- + a1nxn + xn+1
= b1
a21x1 + a22x2 + ---------------------- + a2nxn
+ xn+2
= b2
---------------------------------------------------------------------am1x1 + am2x2 + ---------------------- + amnxn
+ xn+m
= bm
and x1 0 , x2 0 ,------- , xn 0 , xn+1 0 ------ , xn+m 0

Step 1. Basic feasible solution is x1 = x2 = x3 = ------- = xn = 0 (non-basic variables)
xn+1 = b1 , xn+2 = b2 , ------------- , xn+m = bm, (basic variables)
Step 2. Now construct the starting simplex table
cj
c1
c2
cn
Basic
CB
Variables
xn+1
0
xn+2
0
XB
x1
b1
b2
a11
a21
a12
a22
xn+m
bm
am1
am2
amn
z = CBXB
x2 ------------- xn
a1n
a2n
xn+1
xn+2
1
0
0
1
0
------------
xn+m
MIN RATIO
0
0
0
1
0
j = CBXj - cj
Step 3. Incoming vector is considered corresponding to most ve value of j i.e. k

Step 4. Optimality test
a)
If all j 0 , the solution under test will be optimal.
b)
If atleast one j is negative , the solution under test is not optimal then proceed to improve
the solution in the next step.
c)
If corresponding to any negative j , all elements of the column Xj 0 then the solution
under test will be unbounded.
Step 5. Compute min ratio XB / xk for xk > 0

Step 6. Outgoing vector is minimum of min ratio.
Step 7. Element at the intersection of minimum min ratio and incoming vector is key element. Now divide all
the elements of row by key element and make other elements of key column zero. Then we get revised
simplex table and go to step 4.
208
Example 1. Solve by simplex method

Min. -z = a1 - 3a2 + 2a3 subject to
3a1 - a2 + 3a3 7,
-2a1 + 4a2 12,
-4a1 +3a2 + 8a3 10
a1, a2, a3 0
Sol. The problem is of minimization. Converting the objective function from minimization to maximization,
we have
Max. -z = -a1 + 3a2 - 2a3 = Max. z, where z = z
3a1 - a2 + 3a3 + a4 = 7,
-2a1 + 4a2 + a5 =12,
-4a1 +3a2 + 8a3 + a6 = 10
Basic
Variables
a4
a5
a6
a1 = a2 = a3
CB
XB
a1
a2
a3
a4
a5
a6
Min. Ratio (XB / ak)
0
7
0
12
0
10
z =0, z = 0
3
-2
-4
1
-1
4
3
-3*
3
0
8
2
1
0
0
0
0
1
0
0
0
0
1
0
--12/4
10/3
j
5/2
-1/2
-5/2
1/2*
0
1
0
0
3
0
8
2
1
0
0
0
-3/4
0
0
1
0
10 / 5/2
----j
1
0
0
0
0
1
0
0
1/10
3/10
-1/2
8/10
0
0
1
0
j 0
=0
a4
a2
a6
a1 = a2 = a3
0
3
0
10
3
1
Z=9,
=0
z = -9
a1
a2
a6
a3 = a4 = a2
-1
4
3
5
0
11
z=11, z=-11
6/5 2/5
3/5 1/5
11
1
13/5 1/5
=0
The optimal solution is x1 = 4 , x2 = 5, x3 = 0, Min z = -11
9.10.1 Limitations
There are some limitations also with linear programming in spite of wide usage
i.
Parameters used in the model are assumed to be constant, but in real life situation, they are neither
constant nor deterministic.
ii. There are some problems in real life situation where objective functions and constraints are not
linear. Concerning business and industrial problems, constraints are not treated linearly.
iii. Integer valued solutions are not guaranteed. For example, if calculation of men & machine is required
to do some particular job, rounding off the solution to the nearest integer will not give an optimal
solution.
209
iv. Sometimes it is not possible to sort out large scale problems in linear programming in spite of
availability of computers. The problem can be solved by dividing it into lot of small problems and
then sorting out those problems.
v. In Linear programming problems, effect of time and uncertainty are not taken into consideration.
vi. Linear programming deals with only single objective, whereas in real life situations, problems are
having multiple objectives.
9.10.2 Advantages
Following are the advantages of linear programming problem.
i.
The losses occurring due to bottlenecks can be reduced effectively by implementation of linear
programming technique. In production processes, highlighting bottlenecks is one of the most significant
advantage of this technique. For example, when bottlenecks occur, some machines cannot meet
demand while others remain idle for some times.
ii.
Helpful in decision making. By the use of decision making techniques, the user becomes more
objective and less subjective.
iii.
Whenever applicable, linear programming technique helps us in making the optimum utilization of
production resources. It also indicates how a decision maker can employ his productive factors
most effectively by choosing and allocating these resources.
iv.
Linear programming techniques provide practically applicable solutions since there may be other
constraints operating outside the problem which must also be taken into consideration.
9.11 Transportation Problems

It is a process of transportation of various amounts of a single homogeneous commodity which are initially
stored at various locations to different destinations in such a manner that cost of transportation is minimized.
e.g., a pharmaceutical company which has lot of plants at different locations. The company has lot of retail
dealers at different locations in different cities of the country. By the use of transportation model / technique,
transportation schedules can be made which can reduce the cost of transportation to minimum.
9.11.1 Mathematical Formulation
Mathematical Formulation can be explained as
Let us take few assumptions m origins (production unit), ith origin processes ai units of a certain products,
n no. of destinations (which may or may not be equal to m) with destination j requiring bj units. Cost of
transportation from different sources (production units) to different destinations (retail dealers), directly or
indirectly can be calculated depending upon distances, shipping hours etc. cij be the cost of shipping xij unit
product from ith origin jth destination.
We also assume that total availabilities
ai at production units satisfy the total requirements bj at
destinations i.e.
ai = bj (i = 1, 2, 3, ......., m; j = 1, 2, ....., n)
ai bj , manipulation can be done to make it to ai = bj)
Now, the problem is to find out, non-negative ( ) values of xij satisfying both, the availability constraints
(In any case, where
210
n
xij = ai
for (i = 1, 2, 3, ......., m;)
...........(A)
j=1
as well as the requirement constraints
m
xij = bj
i=1
for (j = 1, 2, ....., n)
...........(B)
and minimizing the total cost of transportation

m
n
z =
xij cij
i=1 j=1
(objective function)
...........(C)
It may be seen from the equations that constraints equations (A), (B) & the objective function (C) are all
linear in xij so it may be considered as linear programming problem.
9.11.2 Matrix Form
If we take, the transportation problem represented as mathematically formulated above, the set of constraints
constraints
n
xij = ai (i = 1, 2, 3, ......., m;)
j=1
m
and
xij = bj
for (j = 1, 2, ....., n)
i=1
represent m+n equations in mn non-negative variables xij. Each variable xij appears exactly in two constraints,
one associated with the ith origin Oi and other with jth destination Dj. In the above ordering of constraints,
first we write the origin equations and then destination equations. Then the transportation problems can be
stated in matrix form as follows
Minimize z = CX, X Rmn, subject to the constraints AX = b, x 0, b Rm+n
Where x = [x11, x12 ... x1n, x21,......, x2n,...... xm1,...... ,xmn.], C is the cost factor, b = [ a1, a2, .....am , b1, b2, .....bn ) and A is
an (m+n) x mn real matrix containing the coefficient of constraints.
It is worth noting that elements of A are either 0 or 1. Thus the general LPP can be reduced to
transportation problem if
(i) aij s are restricted to values zero and one and(0 & 1);
(ii) The units among the constraints are homogeneous.
For e.g., if m=2 and n=3, the matrix A is given by
1
A
0
1
0
0
1
0
0
1
0
1
0
0
0
1
0
1
1
0
0
211
0
1
0
1
0
0
1
0
0
1
e(1)23 e(1)23
=
I3
I3
Therefore, for a general transportation problem, we may write

A
e(1)mn e(2)mn ...... e(m)mn
In
In
....... In
Where e(i)mn is an m x n matrix having a row of unit elements as its ith row and 0s everywhere else and In is
the n x n matrix
If aij denotes column vector of A associated with any variable xij, then it can be easily verified that
aij = ei + em+j where ei, em+j R are unit vectors
9.11.3 Few Definitions
The terms feasible solution, basic feasible solution and optimum solution may be formally defined with
reference to the transportation problem (T.P.) as follows
i. Feasible Solution: It is a set of non-negative individual allocations (xij) which simultaneously removes
deficiencies, called as feasible solution.
ii. Basic Feasible Solution : A feasible solution to m-origin, n-destinations problem is basic solution if
the no. of positive allocations are m+n-1 i.e., one less than the sum of rows and columns (which
satisfies the non-negativity restrictions x1 0, x2 0,.... xn 0, xn+1 0.... xn+m-1 0
a. Degenerate BFS: A non-degenerate basic feasible solution is the basic feasible solution which has
less than m+n-1 positive allocations xi (i.e. i < 1,2,3........,m+n-1).
b. Non-Degenerate: BFS A basic feasible solution is called degenerate, if condition referred in
Degenerate BFS is not satisfied.
iii. Optimum Basic Feasible Solution: A basic feasible solution is said to be optimum if it minimizes the
total transportation cost.
9.12 Tabular Representation

Let us suppose that there are m factories and n warehouses, the transportation problem can be shown in a
tabular form as below
W1
W2
......
Wj
.......
Wn
Factory capacities
F1
c11
c12
......
c1j
......
c1n
a1
F2
c21
c22
......
c2j
......
c2n
a2
cin
ai
am
Warehouse
Warehouse
actory
Factory
......
Fi
ci1
ci2
......
......
cij
......
Fm
Warehouse
Warehouse
r e quir e me n ts
requirements
......
......
cm1
cm2
......
cmj
......
cmn
b1
b2
......
Bj
......
Bn
212
m
n
ai = bj
i=1
j=1
W1
W2
......
Wj
.......
Wn
Factory
capacities
F1
x11
x12
......
x1j
......
x1n
a1
F2
x21
x22
......
x2j
......
x2n
a2
xin
ai
am
Warehouse
Warehouse
Factory
actory
......
Fi
xi1
xi2
......
......
xij
......
Fm
Warehouse
Warehouse
r e quir e me n ts
requirements
......
......
xm1
Xm2
......
xmj
......
xmn
b1
b2
......
b2
......
b2
m
n
ai = bj
i=1
j=1
Two table shown above can be combined by inserting each unit of product cost cij with the corresponding
amount xij into the cell ( i, j ).
The product of cij & (xij) gives us the net cost of transporting xij units from Factory Fi to Warehouse Wj.
9.13 Solution of Transportation Problem

There are several methods for solution of Transportation Problem given below
North-west corner method

The method may be stated as follows
a. Start from the upper left corner of the requirements table i.e. the transportation matrix framed in
step I and compare the supply of plant I ( call it S1) with the requirement of distribution centre 1 (call
it D1)
i. If D1 < S1 , if the amount required at D1 is less than the number of units available at S1 , set x11
equal to D1 , find the balance supply and demand and proceed to cell (1,2) (i.e., proceed
horizontally)
Distribution centres
Plants
Requirement
1
2
3
1
2 (6)
1 (1)
5
7/1/0
2
3
0 (0)
8 (5)
5/0
3
11
6
15 (3)
3/0
4
7
1
9 (2)
Supply
6/0
1/0
10/5/2/0
2/0
i. If D1 = S1 , set x11 equal to D1 , compute the balance supply and demand and proceed to cell
(2,2) (i.e. proceed diagonally), also make a zero allocation to the least cost cell in S1 / D1
ii. If D1 > S1 , set x11 equal to S1 , compute the balance supply and demand and proceed to cell
(2,1) (i.e. proceed vertically).
213
b.
We continue in this manner, step by step, away from the north-west corner until, finally, a value is
reached in the south-east corner.
We proceed as follows in this particular example
i. Set x11 , equal to 6, namely, the smaller of the amounts available at S1 (6) and that needed at D1 (7)
and
ii. We proceed to cell (2, 1) (rule c). Compare the number of units available at S2 (namely 1) with
the amount required at D1 (1) and accordingly set x21 = 1. Also set x22 = as per rule (b) above.
iii. We proceed to cell (3,2) (rule b). Now supply from plant S3 is 10 units while the demand for D2
is 5 units; accordingly set x32 equal to 5.
iv. We proceed to cell (3,3) (rule a) and allocate 3 there.
v. We proceed to cell (3,4) (rule a) and allocate 2 there.
It can be easily seen that the proposal solution is a feasible solution since all supply and demand
constraints are fully satisfied.
The following points may be noted in connection with this method
i. The quantities allocated are put in parenthesis and they represent the values of the corresponding
decision variables. These cells are called basic or allocated or occupied or loaded cells. Cells
without allocations are called non-basic or vacant or empty or unoccupied or unloaded cells.
Values of the corresponding variables are all zero in these cells.
ii. This method of allocation does not take into account the transportation cost and, therefore, may not
yield a good (most economical) initial solution. The transportation cost associated with this solution
is
Z = Rs. [2 x 6 + 1 x 1 + 8 x 5 + 15 x 3 + 9 x 2] x 100 = Rs. 11,600
Vogels Approximation Method (VAM)
In the transportation matrix if an allocation is made in the second lowest cost cell instead of the lowest,
then this allocation will have associated with it a penalty corresponding to the difference of these two costs
due to loss of advantage. That is to say, if we compute the difference between the two lowest costs for
each row and column, we find the opportunity cost relevant to each row and column. It would be most
economical to make allocation against the row or column with the highest opportunity cost. For a given
row or column, the allocation should obviously be made in the least cost cell of that row or column.
Substep I Write down the cost matrix as below
Plants
Requirement
1
2
3
1
2
1
5
7
[1]
2
3
0
8
5
[3]
214
3
11
6
15 (3)
3
[5]
4
7
1 (1)
9 (2)
2
[6]
Supply
6[1]
1/0[1]
10[3]
Enter the difference between the smallest and second smallest element in each column below the
corresponding column and the difference between the smallest and second smallest
Element in each row to the right of the row. Put these numbers in brackets as shown. For example in column
1, the two lowest elements are 1 & 2 and their difference is 1 which is entered as [1] below column 1.
Similarly, the two smallest elements in row 2 are 0 and 1 and their difference 1 is entered as [1] to the right
of row 2. A row or column difference can be thought of a penalty for making allocation in second smallest
cost cell instead of smallest cost cell. In other words this difference indicates the unit penalty incurred by
failing to make an allocation to the smallest cost cell in that row or column. In case the smallest and the
second smallest elements in a row / column are equal , the penalty should be taken as zero.
Substep 2 Select the row or column with the greatest difference and allocate as much as possible within the
restriction of the conditions to the lowest cost cell in the row or column selected.
In case of tie among the highest penalties, select the row or column having minimum cost. In case oftie in the
minimum cost also, select the cell which can have maximum allocation. If there is tie among maximum
allocation cells also, select the cell arbitrarily for allocation. Following these rules yields the best possible
initial basic feasible solution and reduce the number of iterations required to reach the optimal solution.
Thus since [6] is the greatest number in brackets, we choose column 4 and allocate as much as possible to
the cell (2,4) as it has the lowest cost 1 in column 4. Since supply is 1 while the requirement is 2, maximum
allocation is (1).
Substep III Cross out of the row or column completely satisfied by the allocation just made. For the
assignment just made at (2,4), supply of plant 2 is completely satisfied. So, row 2 is crossed out and the
shrunken matrix is written below
1
2
5
1
3
3 (5)
8
7
11
15
5/0
7
9
3
6/1[1]
10[3]
1
[3]
[5]
[4]
[2]
This matrix consists of the rows and columns where allocations have not yet been made including revised
row and column totals which take the already made allocation into account.
Substep IV Repeat steps 1 to 3 until assignments have been made.
a) Column 2 exhibits the greatest difference of [5]. Therefore, we allocate (5) units to cell (1,2) , since
it has the smallest transportation cost in column 2. Since requirements of column 2 are completely
satisfied, this column is crossed out and the reduced matrix is written against as the table below
Supply
Plants
1
3
1
2 (1)
5
3
11
15
7/6
[3]
[4]
[2]
215
4
7
9
1/0 [5]
10[4]
b) Differences are recalculated. The maximum difference is [5]. Therefore, we allocate (1) to the cell
(1,1) since it has the lowest cost in row 1. Since requirements of row 1 are fully satisfied, it is
crossed out and reduced matrix is given below.
In table below, it is possible to find row differences but it is not possible to find column differences.
Therefore, remaining allocations in this table are made by following the least cost method.
Plants
6/0
15
3/0
10/4/3/0
1/0
c) As cell (3,1) has the lowest cost 5, maximum possible allocation of (6) is made here. Likewise, next
allocation of (1) is made in cell (3,4) and (3) in cell (3,3) as shown
All allocations made during the above procedure are shown below in the allocation matrix.
Supply
Plants
1
2
3
2 (1)
1
5 (6)
3 (5)
0
8
7
[1]
[3]
[3]
5
[3]
[5]
11
6
15 (3)
3
[5]
[4]
[4]
7
1 (1)
9 (1)
6/1/0
1/0
10/4/3/0
[1][1][5]
[1]
[3][3][4]
2/1
[6]
[2]
[2]
The cost of transportation associated with the above solution is

Z = Rs. (2 x 1 + 3 x 5 + 1 x 1 + 5 x 6 + 15 x 3 + 9 x 1) x 100
= Rs. (3 + 15 + 1 + 30 + 45 + 9) x 100 = Rs. 10, 200
Which is evidently least of all the value of transportation cost found by different methods? Since Vogels
approximation method results in the most economical initial feasible solution, we shall use this method for
finding such a solution for all transportation problems henceforth.
9.14 Summary
This unit provides you a detailed description about linear programming problem and transportation problem,
its assumptions, advantages and disadvantages. it also describes various methods for solving linear
programming problem and transportation problem.
Linear programming is a major innovation since World War II in the field of business decision making under
condition of certainty. Linear means that the relationships are represented by straight lines and the programming
means taking decisions systematically. Thus Linear programming is a decision making technique under given
constraints on the assumptions that the relationships amongst the variables representing different phenomenon
216
happen to be linear. We describe two methods for solving linear programming problem i.e. Graphical
method and Simplex method. Graphical method is applicable only to problems having two variables. But
Simplex method is used for problems having two variables as well as more than two variables. Some linear
programming problems involving several product sources and several destinations of products. Such problems
are known as Transportation problems. A common feature of such problems is that all the units available
must be assigned. The method applicable to such problems is known as Transportation method. This
method is used to optimize the distribution costs.
9.15 Key Words
Linear Programming: A mathematical technique for finding the optimal uses of an organization
resources. In mathematical expression, Equation of the form defined as a1 a1+ a2 a2+.............+ an an. is
LP.
Objective: The ultimate goal of the organization.
Inequality: Mathematical expression indicating that minimum/maximum requirements must be met..
Constraint: A limitation on the availability of resources.
Slack variables: We have to add something positive to the left hand side in order to make it an
equality if a constraint has sign.
Surplus variables: We have to subtract something positive to the left hand side in order to make
it an equality if a constraint has sign.
Objective Function: Expression showing the relationship between the variables in the given problem
and the organization.
Feasible Region: The area containing all the possible solutions which satisfy all the constraints in
the problems.
Infeasibility: The condition when there is no solution which satisfy all the constraints in the problem.
Simplex Method: A systematic and an efficient method for solving a LPP.

1.
What is Linear Programming? Explain applications in various areas.
2.
Explain different assumptions and limitations in LP?
3.
Explain
4.
General formulation of LP
Standard formulation of LP
Matrix formulation of LP
XYZ foods company is developing a low calories high protein diet supplement called Hi-Pro. The
specification have been established by a panel of experts.These specifications given in table
217
Units of Nutritional Elements per gm Serving of Basic Foods
Nutritional
Elements
Calories
Protein
Vitamin B
Vitamin A
1
300
200
125
50
Basic Foods
2
200
150
200
135
3
150
100
140
230
Hi-Pro.
Specifications
400
300
200
150
Formulate the LPP to minimize cost.

5.
Solve the following LPP by graphical method & Simplex method

a) Max z = 3x1 + 2x2 s.t. -2x1 + 3x2 9 , x1 5x2 -20 ; x1 , x2 0
[Unbounded solution]
b) Max. z = 120x + 100y s.t. 10x+ 5y 80, 6x + 6y 66 , 4x + 8y 24,
5x + 6y 90 ; x,y 0
[ x= 500 , y = 600 , Max. z = 1200]
c) Min. z = 5x + 3y s.t. x + y 2, 5x + 2y 10, 3x + 8y 12; x,y 0
[ x = 2 , y = 0 , z = 10]
9.17 References
Operations Research by S.D.Sharma, Kedar Nathram Nath Meerut Delhi.
218
Unit - 10 : Networks
Unit Structure:
10.0
10.1
10.2
10.3
10.4
10.5
10.6
10.7
10.8
10.9
10.10
10.11
10.12
Objectives
Introduction
Project Characteristics
Project Management Phases
CPM & PERT Development
Labelling by Fulkersons I-J rule
Critical Path Analysis
Floats and Slack times
PERT
Summary
Key Words
References
10.0 Objectives
Meaning of Project
Characteristics of a project
Project Management Phases
CPM and PERT development
Presentation of network diagram
Rules regarding presentation of network diagram
Critical Path Method for computation of scheduling
PERT Method and its advantages and disadvantages
10.1 Introduction
A project is defined as a set of inter related activities which should be completed in a certain order to
complete the entire project. These activities may or may not be dependent on each other and these have
certain logical sequence as one cannot be started unless some other activities are completed. An activity in
project viewed as job requiring time and resources for its completion. Until recently, planning was seldom
used in the design phase but after technological development took place at a very rapid speed and designing
became more complex with more inter departmental dependence & interaction so there was needed planning
in the development phase.
10.2 Project Characteristics

Few of the important characteristics of a project are:
i. It is not a permanent entity, it is usually a non-repetitive task.
219
ii. It involves heavy investments.

iii. Its end results are pre decided and objectives very much clear.
iv. It can be of very long duration also, six months to 5 years or may be longer up to 10 years also.
v. It can be broken down into identifiable activities which need resources and certain time period for
completion.
vi. There is a schedule for its completion with target date.
10.3 Project Management Phases

Project management may be divided into major three phases
a.
Planning of Project
The planning phase of a project starts with splitting the project into smaller projects, these smaller
projects is further divided into activities, and analysis of these activities by different departments
takes place. The relationship of these activities with respect to each other is defined and corresponding
authorities and responsibilities are stated. In this manner, possibility of overlooking of any activity
necessary for completion is reduced.
b.
Scheduling of Project
The main objective of the scheduling phase is to decide time completion chart for each activity (start
& finish time) and relation between their time completion. Also critical activities are decided and
these activities should be given proper attention so that the project can be completed within stipulated
time period. For non-critical activities, slack / float times must be shown so that they can be used
advantageously when such activities are delayed or limited resources are to be utilized effectively.
c.
Resources Allocation
To achieve the desired & predefined objectives, resources should be allocated effectively. Resources
like manpower, machines, equipments, finance, space etc. are limited in amount, so these should be
effecting the limitation on time of project.
When resources are limited, they may be demanded simultaneously by different users, therefore
systematic allocation of resources becomes urgent for effective use of resources to complete the
project. Choice of resource usually incurs a compromise and the choice of this compromise depends
on the judgment of managers.
d.
Controlling of Project
It is the final phase of project management. By applying scientific methods of management in the
application areas, critical operations are controlled / completed within time completion targets. By
identifying key result areas, their progress through various activity progress charts and activity
completion charts, a better financial as well as technical control of the project is done.
10.4 CPM & PERT Development

In the years 1956~1958, CPM & PERT techniques were developed almost simultaneously. CPM technique
was developed by Walker initially to solve project scheduling company problems and it was extended to
more advanced status by Mauchly Associates. In simultaneous time period, PERT was developed by engineers
220
of US Navy working on Missile programme. It was a major project involving a lot of departments with lot
of interrelated activities with unknown time frames but project was to be completed in the stipulated time
period. To coordinate the activities of various departments, the group used PERT.
Both the techniques are very much similar, network oriented, using the same principle. These are time
oriented method to determine the time schedule of the project. The significant difference between the two
approaches is that the time estimates in CPM are assumed to be more deterministic while in case of PERT,
these are described probabilistically. Now, these days CPM & PERT comprise actually one technique and
difference, if any are only historical. Therefore, these techniques are referred to as project scheduling
techniques.
These techniques may be applied in a lot of areas
Construction of a dam, oil refinery, coal mining etc.

Maintenance of industries, electrification etc.
Designs of supersonic planes, space ships etc.
Cost control.
Designing a prototype of machine.
10.4.1 Few Definitions

It is important to understand few definitions for network presentation. These are
Activity An operation which utilizes resources and consumes time has an end and a beginning is
called an activity. An arrow is used to represent an activity in which its head shows the direction of
progress of project. These are classified into four following categories
a. Predecessor activity: Activities that must be completed immediately prior to start of another
activity.
b. Successor activity: Activities that cannot be started until one or more other activities are
completed but immediately succeed them are called successor activities.
c. Concurrent activity: Activities that can be accomplished concurrently are concurrent
activities. It may be noted that the activity can be a predecessor or a successor to an event
or it may be concurrent with one or more of other activities.
d. Dummy activity: an activity that does not consume any kind of resource but merely depicts
the technological dependence is called a dummy activity.
Dummy activity is inserted in the network in following two situations
a. To make activities with common starting and finishing points distinguishable.
b. To identify and maintain the proper precedence relationship between activities which are
not connected by events.
10.4.2 Rules for drawing Network Diagram
Following are the rules for drawing network diagram.
Rule no. 1 One arrow is used to represent one activity only in the network. One activity can be shown
twice in the network.
Rule no. 2 No two activities can be identified by same end events.
221
Rule no. 3 For any activity A to add in a network, following questions must be asked to ensure right
precedence relationship in the network
What activity can be finished before activity A is being started?
Which activity will follow activity X?
What are the activities which can occur simultaneously with this activity?
10.4.3 Errors
There are mainly three types of errors as explained below
Dangling When there is disconnection of any activity before all other activities are completed is called
dangling.
-----Dangling-----//
10
//
6
Dangling error in network diagram
For e.g. activities (5-7) and (8-9) are not the last activities in the above network diagram. So the diagram
is wrong and shows the error of dangling.
Looping (or cycling) Looping error is also known as cycling error in a network diagram. When cycle in a
network diagram is endless, the error is called an error of looping.
10
Looping
8
//
11
//
12
Redundancy If dummy activity is inserted unnecessarily in a network diagram, is known as the error of
redundancy.
222
//
Redundancy
//
Redundancy Error diagram
10.4.4 Presentation
While doing scheduling of any project with the help of CPM / PERT techniques, the first step is to make a
diagrammatical layout of activities with precedence relationships among activities.
To understand above terms, we can take an example
Considering A & B concurrent activities, C is dependent on A and D is dependent on A & B both. Such a
situation can be handled by Dummy activity.
Dummy activity
3
Event An event may be regarded as a point in time which signifies the completion of some activities
and beginning of the new ones. This is usually represented by a circle O in a network and can be
named as node or connector also. There are three types of events defined
a. Merge Event: When there is more than one activity which is joining an event.
b. Burst Event: When there are two or more than two activities leaving an event.
c. Merge & Burst event: An activity may be a merge and burst event at the same time as
with respect to some activities it can be a merge event and with respect to other set
activities, it may be a burst event.
Merge Event
Burst Event
223
Merge & Burst Event
An activity is the actual performance of a task which requires time and resources for its completion, it is
represented by an arrow and time period of activity is represented on the arrow whereas an event is that
particular instant of time at which some specific part of a project has been or is to be achieved, it is
represented by a circle and numerical in the circle to show particular instance.
Examples of an event Document made, Design completed, Nail fixed
Examples of an activity Assembly of a product, preparing schedule, mixing of concrete, sand etc.
i
Start event
Activity
End event
Sequencing A network is developed by adjoining activities, events and their sequencings. The first
step in the development is to make a precedence relationships, Following points can be referred
a. To find out job to precede
b. To find out concurrent running job / jobs
c. To find out jobs to follow
d. To find out controls in the start and at the end
Few dependency relationships are given below for the construction of a network diagram, the alphabets
stand for activities
i.
A and B control F; C and B control G;
ii. A and B control F; B controls G while C controls G & H.

iii. A controls F & G; B controls G while C controls G & H.
iv. F & G are controlled by A; G & H are controlled by B with H controlled by B & C.
v. A controls F,G & H; B controls G & H with H controlled by C
Sol. The above dependency relationships can be depicted by figures as given below, the dummy activities
are used to find the right solution
(i)
224
(ii)
(iii)
(iv)
A
... (v)
Example 1. New type of water motor is to be designed for a company. Major specifications are provided
and control of project is required, we can divide the project into major activities
Drawings preparation and approval
Cost analysis
Tooling feasibility report
Tools manufacturing
Favorable cost
Raw materials procurement
Sub assembly order
225
Receiving of sub assemblies

Manufacturing of parts
Final assembly
Testing & shipment
We can make an activity precedence relationship table by using the above information
Activity
Description
Preceding Activity
Drawings preparation and approval
--
Cost analysis
Tooling feasibility report
Tools manufacturing
Favorable cost
B,C
Raw materials procurement
D,E
Sub assembly order
Receiving of sub assemblies
Manufacturing of parts
D,F
Final assembly
H,I
Testing & shipment
The above data can be used for making network diagram.

Step I Activity A has no preceding activity and is represented by arrowed line.
A
Step II Activities B & C are done concurrently and are preceded by activity A.
B
A
C
Step III : Activity D can be sequenced after C and activity E to be scheduled after B & C. Also activity
Cost favorable cost doesnt consume any resource and time, therefore can be considered as Dummy
activity which is shown by dotted line Remaining steps discussed initially.
226
...
E
Dummy Activity
I
D
10.5 Labelling by Fulkersons I-J Rule

Networks are given labeling in a proper manner for convenience, to understand it very fast, to show systematic
presentation. A procedure was developed for that which is known as I-J rule developed by D.R. Fulkerson.
Main steps of the procedure are
1.
Network starts with a starting event which has arrows emerging out of it and none of the arrow
entering into it. Starting event can be given no.1.
2.
Delete all arrows from all numbered events. This will create at least one new start event out of
preceding events.
3.
Number all new start events 2, 3 and so on but numbering from top to bottom may facilitate
other users in reading the network when there are more than one new start events.
4.
Steps 2 & 3 are repeated till the end of network reaches.
The network shown below...
Dummy Activity
J
I
..... can be applied Fulkersons 1-J rule to give nodes numbering
227
By using Fulkersons 1-J rule in above network, numbering of nodes1 & 2 is palpable. By applying step
2, bottom node is the only node through which job is emerging but none entering from the top one. This is
node 3. Applying step 2 again, node number 4 & 5 are easily obtained. Using step 2 again, either of the
node could be numbered 6, as there are two starting points. 4, 5 & 6 being in the same row, top node can
be numbered as 6.
Remaining network is simple and can be made by applying previous discussed rules.
10.6 Critical Path Analysis

After constructing network, time planning becomes essential for planning different activities and hence the
project. An activity-time analysis for different activities is the time duration which activities take for completion.
In network, length of an arrow is not directly proportional to the time taken by it.
This is called as schedule planning and should include following factors
i. Total completion time of project.
ii. Earliest time in starting an activity
iii. Latest time to start the activity without delay in duration of completion of project.
iv. Float of every activity (the time duration for which an activity can be delayed without delaying the
completion of project)
v. Identifying critical activities and critical path.
10.6.1 Basic notations for computations
TE / Ei Earliest occurrence time of event i
TL / Lj Latest allowable occurrence time of event j
Dij Estimated completion time of activity (i,j)
(Es)ij Earliest starting time of activity (i,j)
(Ef)ij Earliest finishing time of activity (i,j)
(Ls)ij Latest starting time of activity (i,j)
(Lf)ij Latest finishing time of activity (i,j)
(i,j) Activity (i,j) with tail event i and head event j
Forward Pass computations / calculations (for Earliest event time) We can find out earliest start and
earliest finish time for each activity (i,j) by these computations.
Step 1 It starts with start node and moves to end node with assumption of earliest occurrence time
of zero for earliest event.
Step 2
Earliest start time of activity (i,j) is the earliest start time of previous activity + time of that activity
(Es)ij = Ei + Dij
Earliest finish time of activity (i,j) is the earliest start time + the activity time
(Ef)ij = (Es)ij + Dij or (Ef)ij = Ei + Dij
228
Earliest event for event j is the maximum of the earliest finish time of all activities into that event
(Ef)ij = (Es)ij + Dij or (Ef)ij = Ei + Dij
Backward Pass computations / calculations (for Latest Allowable time) The latest times (L) indicates
the time by which all activities entering into that event must be completed without delaying the completion of
project.
Step 1 We assume E = L for ending event.
Step 2 Latest finish time for activity (i,j) is equal to the latest event time of event j, i.e. (Lf)ij = Lj
Step 3 Latest starting time for activity (i,j) = the latest completion time of (i,j) activity time
or
(Ls)ij = (Lf)ij - Dij
or
(Ls)ij = Lj - Dij
Step 4 Latest event time for event i is the minimum of the latest start time of all activities originating
from that event i.e.
Li = min. [(Ls)ij for all immediate successors of (i,j)]
= min.[(Lf)ij - Dij] = min.[Lj -Dij]
10.7 Floats and Slack Times

Total Float: The time duration by which the completion of an activity could be delayed beyond the earliest
expected completion time without affecting the overall project duration time.
(Tf)ij = Latest start Earliest start for activity (i - j)
(Tf)ij = (Ls)ij - (Es)ij
Free Float: The time duration by which the completion of an activity can be delayed beyond the earliest
finish time without affecting the float of succeeding activities.
(Ff)ij = Ej - Ei - Dij
Independent Float: The time duration by which the start of an activity can be delayed without affecting the
earliest start time of following activities assuming that the preceding activities have finished at its latest finish
time.
(If)ij = Ej - Ei - Dij
Event Slacks: For any given event, the event slack is defined as the difference between the latest event and
earliest event times. Mathematically for a given activity (i,j),
Head event slack = Lj - Ej
Tail event slack = Li - Ei
All the floats can be represented in terms of head and tail event slack also
Total float = (If)ij = Lj - Ei - Dij
Free float = Ej - Ei - Dij = (Lj - Ei - Dij ) (Lj - Ej )
= Total float Head event slack
229
Independent float = Ej - Li - Dij = Ej - Ei - Dij (Li - Ei )

= Free float Tail event slack
Example2. Find out Early start time (TE) and late start (TL) in respect of all node points and draw critical
path in respect of the following network.
Sol.
2
10
1
7
8
16
13
7
12
10
7
4
15
Earliest Start & latest finish time for different tasks

E1 = 0, L1 = 0
E2 = E1 + D12 = 0 + 10 = 10,
E3 = E1 + D13 = 0 + 8 = 8,
E4 = E1 + D14 = 0 + 9 = 9,
E5 = E2 + D25 = 10 + 8 = 18
E6 = E4 + D46 = 9 + 6 = 15
E7 = max.[E5 + D57 , E3 + D37 , E4 + D46 = max.[18 + 7, 8 + 16, 9 + 7] = 25
E8 = E5 + D58 = 18 + 6 = 24
E9 = E6 + D69 = 16 + 5 = 21
E10 = max.[E7 + D7,10 , E8 + D8,10 , E9 + D9,10 = max.[25 + 12, 24 + 13, 21 + 15 ] = 37
2
10
Calculation of Li (Backward)
L10 = E10 = 37
L9 = L10 D9,10 = 37 15 = 22
L8 = L10 D8,10 = 37 13 = 24
230
L7 = L10 D7,10 = 37 12 = 25
L6 = min.[Lj Dij], where j = 7,9
= min.[25 7, 21 - 5] = 1
L5 = min.[Lj Dij], where j = 7,8
= min.[25 7, 24 - 6] = 18
L4 = 25 8 = 17
L3 = 25 16 = 9
L2 = 18 8 = 10
L1 = 25 8 = 17
Activity (i,j)
Normal time
(Di,j)
(1,2)
(1,3)
(1,4)
(2,5)
(4,6)
(3,7)
(5,7)
(6,7)
(5,8)
(6,9)
(7,10)
(8,10)
(9,10)
10
8
9
8
7
16
7
7
6
5
12
13
15
Earliest time
Latest time
Start
Finish
Start
Finish
(Ei)
(Ei + Di,j) (Lj Dij)
(Lj)
0
10
0
10
0
8
1
9
0
9
1
10
10
18
10
18
9
16
10
17
8
24
9
25
18
25
18
25
16
23
18
25
18
24
18
24
16
21
17
22
25
37
25
37
24
37
24
37
21
36
22
37
Float time
(Lj Dij) - Ei
0
1
1
0
1
1
0
2
0
1
0
0
1
Example 3. Draw a network for the simple project of erection of steel works for a shed. The various
activities of the project are as under
Activity
Description
Preceded by
Erection of site workshop
---
Fencing site
---
Bend reinforcement
Digging foundation
Fabrication of steel work
Installation of concrete pillars
Placing reinforcement
C,D
Concrete foundation
G,F
Erection of steel work
Painting of steel work
H,I
Giving finishing touch
231
Sol.
i.
Activities A & B are without preceding activities
ii.
Activities C & E start after A has finished, activities D & F start after B has finished
E
C
A
iii.
G,F are preceding to H and have common event. Similarly H, I activities are
preceding to J and have common event.
I
G
Example 4. Draw a network for the following activities of a project and number the events according to
Fulkersons rule
A is start event and K is the end event
A precedes event B,
J is successor event to F,
C & D are successor event to B,
D is the preceding event to G,
E & F occur after event C,
232
E precedes F,
C restraints the occurrence of G and G precedes H,
H precedes J & K succeeds J,
F restraints the occurrence of H.
Sol. The events A, B, C, D, E, F, G, H, J & K constitute the network
Activity G depends both on C & F, also L depend upon C hence dummy activity D1 is used. Activity I
depends both on H & L, therefore dummy activities D2 & D3 are used Activity N depends both on H & M
hence dummy activity D4 is used.
10.7.1 Limitations
Following are the limitations of CPM
It does not incorporate statistical analysis in determining time estimates.
It operates on the assumption that there is a precise known time that each activity in the project will
take but this may not be a true assumption in real life.
It is difficult to use CPM as controlling device for the reason that one must repeat the entire evaluation
of the project each time changes are introduced into the network.
10.8 PERT
PERT approach takes uncertainty into account which is associated with the completion of job. In this
method / approach, there are three time values are associate
Optimistic value
Pessimistic value
Most likely value
Optimistic time is the shortest possible time, the activity takes for completion. It assumes that everything
goes very well. It is denoted by to.
Most likely time is the estimation of the time that is taken normally by the activity. It assumes normal
delays. It is denoted by tm.
233
Pessimistic time is the maximum time taken by the activity when everything goes (assumes to be) wrong.
It is denoted by tp.
Expected time is the average time an activity will take if the activity were to be repeated on large number
of times and is based on the assumption that the activity time follows Beta distribution. It can be represented
by the formula mentioned below
te = (to + 4tm + tp) / 6
Variance (2) for an activity can be explained by the formula
2 = [(tp - to) / 6]2
Example 7. Network diagram for activities is given in which three time estimates to, tm, tp are given in the
order of (to - tm - tp) Find out variance and expected time for each activity.
20
14 -17 - 20
30
13 18 - 25
6 9 12
7 10 - 13
10
12 15 - 18
35-9
70
40
47-8
50
9 11 -12
60
Sol. Events can be tabulated in the table

Activity i-j
to
tm
tp
= [ tp - to / 6]
te =
(to + 4tm + tp) / 6
Predecessor
Event i
10
Successor
event j
20
12
1.00
9.0
20
50
0.44
6.7
20
30
14
17
20
1.00
17.0
20
40
10
13
1.00
10.0
20
50
1.00
5.33
30
70
13
18
25
4.00
18.33
40
60
10
14
16
1.00
13.67
40
70
12
15
18
1.00
15.00
50
60
11
12
0.25
10.83
60
70
17
20
25
1.78
20.33
234
Example 8. The time estimates (in weeks) for the activities of a PERT network are given below
Activity
1-2
1-3
1-4
2-5
3-5
4-6
5-6
to
1
1
2
1
2
2
3
tm
1
4
2
1
5
5
6
tp
7
7
8
1
14
8
15
a)
Draw the project network and identify all the paths through it.
b)
Determine the expected project length
c)
Calculate the standard deviation and variance of the project length
Sol. The network for the above mentioned data can be drawn as below
Various paths through the network are 1-2-5-6,
1-3-5-6,
And 1-4-6.
2
1-1-7
1-4-7
1-1-1
2-5-14
2-2-8
3-6-15
2-5-8
b) For determining the expected project length, the expected activity times need to be calculated.
Activity i-j
Predecessor
Event i
1
1
1
2
3
4
5
Successor
Event j
2
3
4
5
5
6
6
to
tm
tp
= [ tp - to / 6]
te =
(to + 4tm + tp) / 6
1
1
2
1
2
2
3
1
4
2
1
5
5
6
7
7
8
1
14
8
15
235
1
1
1
0
4
1
4
2
4
3
1
6
5
7
te = expected completion time of each activity

TE = expected completion time of the Project.
Length of path 1-2-5-6
2 + 1 + 7 = 10
Length of path 1-3-5-6
4 + 6 + 7 = 17
Length of path 1-4-6
3+5
=8
Since 1-3-5-6 has the longest duration, it is the critical path of network.
Therefore, expected project length TE = 17 weeks.
c) Variance of the project length is the sum of the variances of the activities on the critical path.
Vcp = V1-3 + V3-5 + V5-6 = 1+4+4 = 9
S.D = cp =
Variance
= 3
10.8.1 Advantages and Disadvantages
Following are the advantages and disadvantages of PERT .
Advantages
It concentrates attention on critical elements that may need correction.
It forces managers to plan.
It forces planning all down the line.
It makes possible a kind of feed forward control. A delay will affect succeeding events unless the
manager can somehow make up the time by shortening that of some actions in future.
It makes possible the pressure for action at right spot and level in the organization structure at the
right time.
Disadvantages
It has emphasis on time only and not on costs.
It is not practicable for routine planning of recurring events.
It is not a cure-all. It will not do the planning although it forces planning.
The calculations of probabilities is done on the assumption that a large number of independent
activities operate on critical path and as such the distribution of total time is normal but this may not
be true in real life which reduces the significance of all probability calculations that we make under
PERT.
It does not consider the resources required at various stages of the project.
If time estimates to perform activities are unsatisfactory then the network diagram and the critical
path will have little real meaning after project begins.
236
10.9 Summary
This unit gives you a detailed description about network analysis. Management of any project involves
planning, coordination and control of a number of interrelated activities with limited resources such as
machine, men, money and time. Furthermore, it becomes necessary to incorporate any change from initial
plan as they occur and immediately know the affect of change. Therefore, the managers are forced to look
for and depend on a dynamic planning and scheduling system which will produce the best possible initial
plan and schedule. The quest for such a technique led to recent development of network analysis.
Network Analysis is a technique related to sequencing problems which are concerned with minimizing some
measure of the system such as the total completion time of project, the overall cost and so on. The technique
is useful for describing the elements in a complex situation for the purpose of designing, planning, coordinating,
controlling and making decision. It is specially suited to projects which are not in routine and conducted only
few times. Two most popular forms of this technique used in many scheduling situations are the Critical Path
Method (CPM) and Programme Evaluation and Review Technique (PERT).
10.10 Key Words
Project: A set of inter related activities which should be completed in a certain order to complete
the entire project.
Fulkersonss 1-J Rule: in Networks the nodes are given labeling the procedure was developed
by D.R. Fulkerson for that which is known as I-J rule
Activity An operation which utilizes resources and consumes time
Predecessor Activity: Activities that must be completed immediately prior to start of another
activity.
Successor Activity: Activities that cannot be started until one or more other activities are completed.
Concurrent Activity: Activities that can be accomplished concurrently
Dummy Activity an activity that does not consume any kind of resource but merely depicts the
technological dependence.
Earliest Start Time: Earliest time at which an activity can begin.
Earliest Finish Time: Earliest time at which an activity can be completed
Event A point of time which signifies the completion of some activities and beginning of the new
ones.
Optimistic Time: is the shortest possible time when everything assumed to be going right, the
activity takes for completion, denoted by to.
Most likely Time: is the estimation of the time that is taken normally by the activity, denoted by tm.
Pessimistic Time: is the maximum time taken by the activity when everything assumed to go
wrong, denoted by to.
Expected Time: is the average time an activity will take if the activity were to be repeated on large
number of times
237
Programme Evaluation & Review Technique (PERT) & Critical Path Method (CPM):
PERT and CPM are two techniques of OR which are used to show diagrammatical representation
of various activities, their relations, dependencies etc.

1.
What is PERT & CPM? Explain with examples.
2.
What is network analysis? When is it used? What is meant by the phrase Critical Path?
3.
Explain different types of Floats.
4.
What are predecessors and successors, explain with examples.
5.
How different types of errors occur in network, explain with diagrams.
6.
A project consists of series of tasks labeled A,B,,I with the following relationship(K
< L,M means L and M can not start until K is completed, L,M < K means K can not start until both
L and M are completed. Construct the network diagram having the following constraints:
A < D,E
B, D < F
C<G
B<H
F, G <I
Also find the optimistic time of completion of the project when the time(in days) of completion of
each task is as follows
Task
Time
1.
A
23
B
8
C
20
D
16
E
24
A project has the following time schedule

Activity
(1-2)
(1-3)
(2-4)
(3-4)
(3-5)
(4-9)
(5-6)
(5-7)
(6-8)
(7-8)
(8-9)
(8-10)
(9-10)
Construct the network and compute
238
F
18
G
19
H
I
4
10
[Ans. 67days]
Time in weeks
4
1
1
1
6
5
4
8
1
2
1
8
7
8.
Find out critical path of the network given below using following data
Activity
Least time
Greatest time
Most likely time
A
4
8
5
b
5
10
7
c
8
12
11
d
2
7
3
e
4
10
7
f
6
15
9
g
8
16
12
h
5
9
6
I
3
7
5
j
5
11
8
k
6
13
9
II
III
IV
VII
VI
VIII
(Ans. a-c-e-h-k)
9.
1.
A project has following details; PERT network is to be constructed. Find also Critical Path.
Activity
1-2
2-3
2-4
3-5
4-5
4-6
5-7
6-7
7-8
7-9
8-10
9-10
Most optimistic time, to

1
1
1
3
2
3
4
6
2
5
1
3
Most pessimistic time, tp

5
3
5
5
4
7
6
8
6
8
3
7
Most likely time, tm

1.5
2
3
4
3
5
5
7
4
6
2
5
(Ans. 1-2-4-6-7-9-10)
10.19 References

239
Unit - 11 : Queuing and Game Theory

Unit Structure:
11.0
11.1
11.2
11.3
11.4
11.5
11.6
11.7
11.8
11.9
11.10
11.11
Objectives
Introduction
Meaning of Queue
Elements of the Queuing System
Assumptions of the Queuing Model
Single-Channel Poisson Arrivals with Exponential Service Rate (M/M/1)
Game Theory
Types of Games
Summary
Key Words
References
11.0 Objectives
Concept of queuing system
Queuing Models(M/M/1, M/M/k)
Concept of Games Theory
Types of Games
11.1 Introduction
Queues are very common in everyday life. We quite often face the problem of long queues for a bus, banks,
barber shop, a movie ticket and for various other situations. In large cities, long queues are seen in front of
railway booking offices, post offices, automobiles waiting at service stations, ships waiting for berths and
patients waiting for doctors. Queues are thus a common phenomenon of modern civilized life. The theory of
queuing models has its origin in the work of A.K.Erlang, a Danish Engineer of the Copenhagen Telephone
Company during 1910s.
Decision situations frequently arise in which units arriving for service must wait before they can be serviced.
If the laws governing arrivals, service times and the order in which the arrival units are taken into service are
known then the nature of the waiting situation can be studied mathematically in a simplistic way.
A queuing problem arises when a current service rate of a facility falls short of the current flow rate of
customers. If a service facility is capable of servicing the customer when he arrives, no bottleneck will occur.
But if it takes 10 minutes to service a customer and one customer arrives every 8 minutes, a queue will build
up to infinite length if same arrival and service rate continues. In such a situation, this bottleneck is eliminated
only if either arrival rate decreases or service rate increases. If size of the queue happens to be a large one
then at some times people leave the queue due to impatience and a sale is lost by the concerned business
unit. Hence the queuing theory is concerned with the decision making process of the business unit.
240
11.2 Meaning of Queue

Ordinarily the forms in front of service facilities are called a waiting line or a queue. A queue thus involves
arriving customers who want to be serviced at the facility which provide the service they want to have. In
short, the word queue refers to waiting in line.
The idea about a queue may be expresses as under
Customers arrival
Waiting line/Queue
Getting serviced
Departure of customers
11.2.1Commonly Used Terminology in Queuing Theory
Queuing System: A system consisting arrival of customers, waiting in queues, picked up for service,
being serviced and the departure of customers.
Customer: Persons arriving at a station for service. Customers may be either persons or other
items.
Service station: Point where service is to be provided.
Queuing length: It is the number of customers waiting in the queue.
Waiting time It is the time a customer spends in the queue before being serviced.
Number of customers in the system: It is the sum of number of customers in the queue and
number of customers being serviced.
Time spent by a customer in the system: It is the sum of waiting time and service time.
Jockeying: Leaving the first queue and joining the other.
Reneging: Joining the queue and leaving it afterwards.
Balking: Customers decides not to join the queue.
11.3 Elements of the Queuing System

A queuing system has the following elements
Arrivals
Service mechanism
Queue discipline
Output of the queue
241
Arrivals
Customers arrive at a service station for service. They do not come at regular intervals but an arrival into the
system occurs according to some chance mechanism. Often arrival occurs at random and is independent of
what has previously occurred. Arrivals may occur at a constant rate or may be in accordance with some
probability distribution such as Poisson distribution, Normal distribution etc. The following information is
considered relevant for input process
i)
The source population

a) Infinite (very large)
b) Finite (limited Number)
ii) Arrival distribution

iii) Inter-arrival distribution
iv) Mean arrival rate i.e. the average number of customers arriving in one unit of time. It is represented
by .
v) Mean time between arrivals i.e. 1/ .
Service mechanism
It is concerned with the service time and the service facilities. Service time can either be fixed or distributed
in accordance with some probability distribution. Service facilities can be of following types.
Single Channel facility: In this, there is only one queue in which the customer waits till the service point is
ready to take him for servicing.
One queue several service stations facilities: In this, customer wait in a single queue until one of the
service stations is ready to take them for servicing.
Several queues one service station: In this, customer can join any one of queue but the service station
is only one.
Multi channel facility: In this, there are many queues and many service stations facilities.
Multi stage channel facility: In this, customers require several types of services and there are different
service stations. The customer has to get service from more than one service stations in some sequence; one
after another. Each station providing a service station before leaving the system.
The following information is considered relevant for service mechanism
i)
Distribution of number of customers serviced.
ii) Distribution of time taken to service customers.

iii) Average number of customers being serviced in one unit of time at a service station. It is represented
by .
iv) Average time taken to service a customer.
Queuing discipline
Specifically it means, existence of some rule according to which the customers actions as to when their turn
comes up for the service. It may be First In First out (FIFO) or Last in Last out (LIFO). Mostly, FIFO rule
is applicable in Queuing systems.
242
Output of the Queue

In a single channel facility, the output of the queue does not pose any problem for the customer departs after
receiving the service but in multi channel facilities, output of queues becomes important.
11.4 Assumptions of the Queuing Model
One queue
One service station
Queue discipline FIFO( First In First Out)
The population from which the queue arise is sufficiently large( assumed to be infinite)
Arrivals and services are coming individually
Arrivals and services occur in accordance with a Poisson process
Time intervals between arrivals and services follow exponential distribution
An overall assumption is that the system has been in operation for some time and has stabilized.
the average arrival rate is less than average service rate i.e. /<1
11.5 Single-Channel Poisson Arrivals with Exponential Service Rate (M/M/1)

A single channel (1), Poisson arrivals (M) with exponential service (M) problem treats a condition where
one unit is delivering the service. The inputs such as customers are considered to arrive in a Poisson
manner. The exponential servicing rate is independent of the number of elements in a line. The following
notations are used in queuing models
mean arrival rate/unit per time
t probability that an arrival enters the system between t and t + t time interval
- mean service rate
t probability of completion of a service between t and t + t time interval
n number of customers in the system at time t
Pn probability that exactly n customers are in the system at time t
Following formulae are used when < 1.
The expected proportion of time a facility will be idle
P0 = 1
(The system is idle when there is no customer/unit in the system).
The probability that the exact number of customers n in the system at a given time
P (n) = ( ) n (1 - )
Expected number of units in the system

E (n) =
243
Expected number of units in the queue is

E (m) =
2
( )
Average waiting time in the queue of an arrival is

E (w) = E (m) /
( )
Average time an arrival spends in the system is

E (v) =
1
( )
The probability that the number in the queue and being serviced is > k is
P (n>k) = (
k+1
)
Average length of non-empty queue

E (m | m > 0) =
( )
Average waiting time of an arrival that waits

E (w | w > 0) =
1
( )
Example 1. Mean arrival rate of a queue is one customer in every 4 minutes and mean service time is
2.5 minutes. Calculate
a)
b)
c)
d)
Sol.
average number of customers in the system

average queue length
average time a customer spends in the system
average time a customer waits before being served.
1
= 0.25 arrivals/minute
4
= 15 arrivals /hour
=
1
= 0.4 service/minute
2.5
= 24 services/hour
15
=
= 0.625 and is less then 1.
24
244
a) Average number of customers in the system
E(n) =
= ( 15/24-15) = 15/9
b) Average queue length
2
( )
E(m) =
0.25 2
= 1.04 customers
0.4(0.4 0.25)
c) Average time a customer spends in the system

E(v) =
1
( )
= 1/ (0.4 0.25) = 6.66 minutes

d) Average time a customer waits before being served
E(w) =
( )
0.25
0.4(0.4 0.25)
= 4.16 minutes
Example 2. In a telephone booth, arrivals are considered to be Poisson with an average time of 10 minutes
between one arrival and the next. The length of a phone call is assumed to be distributed exponentially with
mean 3 minutes. Calculate
a) What is the probability that a person arriving at the booth will have to wait?
b) What is the average length of the queues that form from time to time?
c) The telephone company will install a second booth when convinced that an arrival would expect to
have to wait at least 3 minutes for the phone. By how much should the flow of arrivals be increased
in order to justify a second booth?
Sol. = 0.1 arrivals/minute
= 0.33 service/minute
a) Probability(an arrival has to wait) = 1 P0
=
0. 1
0.33
= 0.3
245
a) Average length of non-empty queues

E(m | m > 0) =
( )
= 0.33/ (0.33 0.1)

= 1.43 persons
b) Average waiting time for an arrival before gets service
E(w) =
( )
E(w) = 3
Given
3 = / 0.33(0.33 - )
= 0.16 arrival/minute
When rate of arrival is 0.16 per minute (9.6 per hour) or more.
Example 3. Machine X break-down at an average rate of 5/hour. The breakdowns follow Poisson process.
Cost of idle machine hour comes to Rs. 15/hour. Two workers A and B have been interviewed- A charges
Rs. 8/hour and he services break-down machines at the rate of 7/hour whereas B charges Rs. 10/hour and
he services at an average rate of 9/hour. Which workers services should be used and why? Assume work
shift is of 8 hours.
Sol. Average breakdown of machines = 5/hour
Expected number of breakdown machines in the system is
E(n) =

Cost of idle machine hour = Rs 15
For worker A
Service rate = 7/hour
Hourly charges = Rs. 8
E(n) =
= 5/ ( 7-5) = 2.5
It means 2.5 machine hours are lost in an hour.
Total machine hours lost in a day = 2.5 X 8 = 20 machine hours.

Total cost = Hire charges of worker + Cost of idle machine
= 8 X 8 + 20 X 15
= 64 + 300 = 364
246
For worker B
Service rate = 9/hour
Hourly charges = Rs. 10
E(n) =
= 5/ ( 9-5) = 1.25
It means 1.25 machine hours are lost in an hour.
Total machine hours lost in a day = 1.25 X 8 = 10 machine hours.
Total cost = Hire charges of worker + Cost of idle machine
= 8 X 10 + 10 X 15
= 80 + 150 = 230
Comparing the total cost for worker A and B; worker Bs service would certainly be more economical.
11.5.1 Multi - Channel Poisson Arrivals with Exponential Service Rate (M/M/k)
In the multi-channel queuing system, there is k number of parallel stations; and each element in the waiting
line can be served by any one station. Each service facility has the same type of service. Single waiting line
is formed, and the customer moves into any one of the service stations (windows) which becomes available
after having serviced a customer. The utilization factor () describes the probability of a given station being
in use.
= Average rate of arrivals.
= Average service rate for each of the service stations.
k = number of service stations.
k= Average combined service rate of all the k service stations.
e=
= utilisation factor for the system.
k
Also
<1
k
1 in a multi-channel system is given by

The probability of having no customers
P0 = _____________________________________
k-1
(1/n!)( )n +[ (1/k!)( )k (k/(k - ))]
n=0
Pn = P0 (1/n!)( )n ------------------------- when n k
= P0(
n
) /K1.Kn-k----------------------- when n > k.
247
The probability that an arrival has to wait

P(n k) =
P0 (/) n
[(k - 1)!(k - )]
Average queue length

E(m) =
P0 (/) n
[(k - 1)!(k - ) 2 ]
Average number of customers in the system

P0 (/) n
E(n) =
+ /
[(k - 1)!(k - ) 2 ]
Average waiting time of an arrival
E(w) = P0 (/)k / [(k1)!(k )2]
Average time an arrival spends in the system
E(v) = P0 (/)k / [(k1)!(k )2] + 1/
Example 4. A firm has 4 service stations to receive people with problems. Arrivals average 80 persons
in an 8-hour service day to have an exponential distribution. The average service time is 20 minutes.
Calculate
a)
b)
c)
d)
e)
f)
g)
Average number of customers in the system

Average number of customers waiting to be serviced
Average time a customer spends in the system
Average waiting time for a customer
How many hours each week does a service station spend performing job?
Probability that a customer has to wait before he gets service.
Expected number of idle service station at any specified time.
Sol. = 10/hour
= 3/hour
k=4
e=
= 10/3 4 = 0.833 <1

k
The probability of having no customers in a system is given by
1
P0 = _____________________________________
k-1
(1/n!)( )n +[ (1/k!)( )k (k/(k - ))]
n=0
248
1
=___________________________________________________________
[1+ + (1/2) * ( )2 + (1/6) * ( )3 + (1/24) * ( )4 (k/(k - ))]
1
=_____________________________________________
1 + 10/3 + 100/18 + 1000/162 + (10000/1944) (12/2)
= 0.0213
(a) Average number of customers in the system.
E(n) = P0 (
) / [(k-1)!(k - )2] +
= 0.0213*10*3(10/3)4 / [3! *(12 10)2] + 10/3

= 6.61 customers
a)
Average number of customers waiting to be serviced

E(m) = P0 (
k
) / [(k-1)!(k - )2]
= 0.0213 * 10 * 3(10/3)4 / [3! *(12 10)2]

= 0.0213 * 154.1
b)
= 3.28 customer
Average time a customer spends in the system
E(v) = P0 (
k
1
) / [(k-1)!(k - )2] +
= 0.0213 * 3(10/3) 4 / [3! *(12 10)2] + 1/3

= 0.661 hour
a) Average time a customer waits before being served (waiting in the queue)
E(w) = P0 (
k
) / [(k-1)!(k - )2]
= 0.0213 * 3 *(10/3)4/[3! *(12 10)2]

= 0.328 hour
b) Total time spent each week by a service station
The utilization factor
= 10/(4*3) = 0.833
249
Expected time spent in servicing customers during an 8-hour day

= 8 X 0.833 = 6.66 hours
Therefore , on an average a service station is busy 33.3 hours based on a 40 hour week.
e) Probability that a customer has to wait
P(n k) = P0 (/)n / [(k1)!(k )]
= 0.0213 *3 * (10/3)4 / [3! *(12 10)2]
= 0.6571
f) Expected number of idle service stations at any specified time
Pn = P0 (1/n!)(
n
)
P 0 = 0.0213
P1 = (1/1!)(10/3)*0.0213 = 0.0709
P2 = (1/2!)(10/3)2 *0.0213 = 0.1182
P3 = (1/3!)(10/3)3 *0.0213 = 0.1314
P0 = 4 Service Stations
Therefore, Expected number of idle service stations at any specified time= 4P0 + 3P1 + 2P2 + 1P3 +0P4
=
4(0.0213) + 3(0.0709) + 2(0.1182) + 1(0.1314) + 0
= 0.666
Thus on average, less than one service station is idle at any time.
11.6 Game Theory

The term game represents a conflict between two or more parties. Game theory is really the science of
conflict. Professor John von Neumann and Oscar Morgenstern published their book entitled The Theory
of Games and Economic Behaviour in which they provided a new approach to many problems involving
conflicting situations. Game theory is not concerned with finding an optimal strategy for a particular conflict
situation but it provides general rules concerning the logic that underlines strategic behaviour of all types.
Game theory applies to competitive situations are known as competitive games. To be termed as games,
situations must possess the following properties.
a) Number of competitors is finite.
b) There is a conflict between the participants.
250
c) Each of the participants has several choices as to his appropriate actions.

d) The rules governing these choices are specified and known to all players.
e) The outcome of the game is affected by choices made by all the players. The choices are to be
made simultaneously so that no competitor knows the opponents choice until he is already committed
to his own.
f) The outcome for all specific sets of choices by all of the players and, known in advance by every
player.
11.6.1 Standard Convention in Game Theory
Game theory has the following set of conventions
a) Omit a description like player A wins five points and replace it with integer 5.
b) Player A has choices between the rows and player B has choices between the columns. Each of the
rows is a strategy for A to choose, similarity each of column is a strategy to choose from.
c) Games are represented in the form of a matrix. When games are expressed in this manner, the
resulting matrix is known as payoff matrix.
d) The term strategy refers to the total pattern of choices employed by any player.
11.7 Types of Games

Games can be of several types given below
Two-person games and n-person games

In it players may have many possible choices for each play of the game but the number of players
remains only two. In case of more than two persons, the game is known as n-person game.
Zero-sum and non-zero sum game

A zero-sum game is one in which the sum of the payments to all competitors is zero for every
possible outcome of the game i.e. sum of points won equals the sum of the points lost. But in a nonzero sum game, sum of the payoffs from any play of the game may be either positive or negative but
not zero.
Games with limited number of moves and unlimited number of moves
In games with limited number of moves, number of moves is limited to fixed magnitude before play
begins but in games with unlimited number of moves, it could be continued over an extended period of
time and no limit is put on the number of moves.
Cooperative and non-cooperative games

Games in which players can negotiate are known as cooperative games and games in which players
can not negotiate are known as non-cooperative games.
2 X 2 Two-person games and 2 X m and m X 2 games

Two person zero-sum games with only two choices open to each player are known as 2 X 2. Two
person games in which one of the players has more than two choices of rows and columns and
other player has exactly two choices is referred as m X 2 or 2 X m game respectively.
251
rows and columns and other player has exactly two choices is referred as m X 2 or 2
X m game respectively.
B
A
-1
-2
-1
-2
Note: The scope of material in this unit is limited to 2 person Zero Sum Games only.
Example 5: Find out value of the game from the following payoff matrix and also determine the optimum
strategies for the two players P1 and P2
Player P2
A1
-1
-2
-2
Player P1 A2
-1
-3
-7
-7
A3
-6
-9
-9
-1
Maximum loss of B
Minimum of Maximum
Sol.
Minimum Gain of A Maximum of Minimum

-2
-1
To determine optimum strategy, we take cautious approach, assume the worst and act accordingly.
The positive values in the matrix are gain to player P1 and loss to player P2.Similarly negative
values in the matrix are loss player P1 and Gain to player P2.
If player P1 plays with first row strategy then player P2 will play with second column for then P2
will win one point otherwise he is to lose 3, 4 and 2 if he plays with column 1, 3 and 4 resp. If player
P1 plays with second row strategy then worst would happen to him only when player P2 plays with
3rd column because in that case P2 wins 7 points. If player P1 plays with 3rd row strategy, then
worst he can expect is of losing 9 points when P2 plays with 4th column. In this problem, then player
P1 should adopt first row strategy because only then his loss will be minimum. Thus player P1 can
make the best of the situation by aiming at the highest of these minimal payoffs. This decision rule is
known as maximin strategy(maximising minimum Gains)
From the point of view of player P2, it can be said that if player P2 plays with column first strategy,
the maximum he can lose is 4 points if player P1 adopts strategy of row 3. If player P2 plays with
column two strategy there is no question of any loss whatever may be the strategy of row one for
only then his loss will be minimum. If player P2 plays with column third strategy the maximum he can
lose is 4 points if P1 adopts strategy of row one. If P2 adopts strategy of column 4 he can lose at the
most 2 points if P1 adopts strategy of row one. In this problem then P2 should adopt second
column strategy and thereby ensure a win of 1 point which is the maximum in the given case. Thus
player P2 can make the best of the situation by aiming at the lowest of these maximum payoffs (viz.
4, -1, 4, 2). Thus he should seek the minimum among the maximum payoffs. This decision rule is
known as Minimax strategy.
252
Player P2
Player P1
-1
-1
-3
-7
-6
-9
-2
-2
Choice of p1
Choice of P2
The Saddle point and Pure Strategy
The saddle point in a payoff matrix is one which is the smallest value in its row and the largest value in its
column. The saddle point is also known as equilibrium point in the theory of games. An element of a matrix
that is simultaneously minimum of the row in which it occurs and the maximum of the column it occurs is a
saddle point of the game. In a game having a saddle point; optimum strategy for player P1 is always to play
the row containing a saddle point and for the player P2 to always play the column that contains a saddle
point. Saddle point also gives the value of such a game. Such a choice is known as PURE STRATEGY i.e.
whenever a player chooses only one strategy for all repetitions of the game; and his gain will be less, if he
shifts to another strategy.
Saddle point in a payoff matrix concerning a game may be there and may not be there, If there is a saddle
point we can easily find out the optimum strategies and the value of the game .Its saddle point does not exist
we have to use algebraic methods for working out the solutions concerning the game.
Example 6: Find the optimum strategies and the value of the game from the following payoff matrix concerning
two-person game:
Player P2
Player P1
Solution: In the given game, there is no saddle point; therefore the players will resort to what is known as
MIXED STRATEGY i.e. player P1 will play each of his rows a certain portion of time and player will play
each of his columns a certain part of the time. The question then is to determine what proportion of the time
a player should spend on his respective rows and columns. This can be done by the use of algebraic method
stated as follows:
Let Q equal the proportion of time player P1 spends playing the first row, then (1-Q) must equal the time he
spends playing his second row (because one equals the time available for play). Similarly, suppose player
P2 spends time P in playing first column and (1-P) proportion of time he spends playing the second column.
All this can be stated as under. Player P2
Player P1
1-P
1-Q
3
253
Now, we must find out the values of Q & P. Let us analyse the situation from P1s point of view. He would
like to devise a strategy that will maximise his winning (or minimise his losses) irrespective of what his
opponent Y does. For this P1 would like to divide his play between his rows in such a manner that his
expected winnings or losses when P2 plays the first column will equal his expected winnings or losses when
P2 plays the second column. (Expected winnings indicated the sum, over time, of the payoffs that will obtain
multiplied by the probabilities that these payoffs will obtain). In our case, we can calculate the same as
under:
P1s EXPECTED WINNINGS
When P2 plays
When P2 plays
Column one
P1 plays row one
Column two
(1) (Q)
(4) (Q)
(5) (1-Q)
(3) (1-Q)
Q of the time
P1 plays row two
1-Q of the time
P1s total expected winnings
Q + 5(1-Q)-----(1)
4Q + 3(1-Q)-----(ii)
Equating the expected winnings of P1 when P2 plays column one when P2 plays column two; we can find
the values of Q as follows:
Q + 5(1-Q) = 4Q + 3(1-Q)
Q + 5 5Q = 4Q + 3 3Q
-5Q = -2
Q=2/5
1- Q = 3 /5
This means that player P1 should play his first row 2/5 of the time and his second row 3/5 of the time if he
wants to maximise his expected winnings from the game.
On the similar basis expected losses of P2 can be worked out as under:
P2s EXPECTED WINNINGS
When P2 plays
P1 plays row one

P1 plays row two
When P2 plays
Column one
Column two
P of the time
1-P of the time
(1) (P)
(5) (P)
P + 4(1-P) = 5P + 3(1-P)
P + 4 4P = 5P + 3 3P
254
P2s total
expected losses
(4) (1-P)
(P)+ 4(1-P)
(3) (1-P)
5R + (3) (1-P)
-5P = -1
P=1/5
1- P = 4 /5
Value of the game = Q+5(1-Q)(i)
Or
4Q + 3(1-Q)(ii)
= 2/5 + 5*3/5
= 3 2/5 Rupees.
Check (ii) also gives the same values.
This means that player P2 should play his first column 1/5 of the time and his second column 4/5 of the time
if he wants to minimise his expected losses in the game.
Now, we can illustrate the original game with the appropriate strategies for each of the player as follows:
Player P2
1/5
Player P1
4/5
2/5
3/5
Algebric Method
Let a 2*2 game be denoted as follows.
P1
P2
Strategy
B1
B2
Q1
A1
a11
a12
Q2
A2
a 21
a22
Price
Step one, Subtract the smaller payoff in each row from the larger one and the smaller
payoff in each column from the larger one.
P2
P1
(i.e. 4-1 = 3)
(i.e. 5-3 = 2)
(i.e. 5-1 = 4)
(i.e. 4-3 = 1)
255
Step 2: Interchange each of these pairs of subtracted numbers found in step 1 above
P2
P1
Step 3: Put each of the interchanged numbers over the sum of the pair of numbers
P2
P1
2/(2+3)
3/(2+3)
1/(1+4) 4/(1+4)
Step 4: Simplify the fraction to obtain the proper proportions or the required strategies
P2
P1
2/5
3/5
1/5
4/5
Now, we determine the values of the game. Looking at the game from P1s point of view we can argue as
follows:
(i) During the 1/5 of the time P2 plays column one P1 wins 1 point 2/5 of the time (when P1 plays row
one) and 5 points 3/5 of the time (when P1 plays row second)
(ii) During the 4/5 of the time P2 plays column two P1 wins 4 points 2/5 of the time (when P1 plays
row one) and 3 points 3/5 of the time (when P1 plays row second)
Thus total expected winnings of player P1 are the sum of the above statements which is under:
1
2
3
4
2
3
{(1)( ) + (5)( )} + {(4)( )+(3)( )}
5
5
5
5
5
5
=(
1 17
4 17
)(
) + ( )(
)
5
5
5
5
17
68
+
25 25
85
25
17
5
256
Thus the value of the game is 17/5 which means that player P1 can expect to win an average payoff of 17/
5 points for each play of the game if he adopts the strategy we have determined as stated above. If the value
of the game determined above had a negative sign, it would simply signify that P2 was the winner. The same
result we can also get by looking at the game from P2s point of view doing similar calculations.
Concept of Dominance to reduce the size of Game to 22
Example 7 . Find the optimal strategies for A and B in the following game. Also obtain the value of the
game.
Bs Strategy
b1
b2
b3
As Strategy a1
-7
a2
-6
a3
-7
The pay-offs in relation to a1 are equal to or greater than the corresponding pay-offs from a 3.
Thus, a1 dominates a3. Deleting a3, we get,
Bs Strategy
b1
As Strategy a1 9
a2 3
b2
b3
-7
-6
Now, we observe that strategy b1 is dominated by b2. Thus, we delete b1. This gives the
following matrix:
Bs Strategy
b1
As Strategy a1
a2 -6
b2
-7
4
Solving 2n or n2 games using graphical method.

Example. Solve the following game using graphical approach:
Bs Strategy
b1
As Strategy a1 8
a2 -6
b2
b3
b4
-7
-2
257
Here A has two strategies a1and a2 which, suppose, he plays with probabilities x and 1-x respectively.
When B chooses to play b1, the expected payoff for A shall be 8x +(-6)(1-x) or 14x-6. Similarly, the
expected pay-off functions in respect of b1, b2, b3 and b4 can be derived as being 6-x; 4-11x; and 11x-2,
respectively. We can represent these by graphically plotting each pay-offs as a function of x. This is shown
in below figure.
The lines are marked b1, b2, b3 and b4 and they represent the respective strategies. For each value of x, the
height of the lines at that point denotes the pay-offs of each of Bs strategies against (x,1-x) for A. A is
concerned with his least pay-off when he plays a particular strategy, which is represented by the lowest of
the four lines at that point, and wishes to choose x as to maximise this minimum pay-off. This is at K in the
figure where the lower envelop (represented by the shaded region), the lowest of the lines at the point, is the
.highest. This point lies at the intersection of the lines representing strategies b1 and b3. The distance KL = 0.4or -2/5) in the figure represents the game value, V, and x = OL(=0.4 or 2/5) is the optimal strategy for
A.
Pay-off
Pay-off
10
10
b2
b1
b4
0.2
0.4
2
0.6
0.8
-2
0
-2
-4
b3
-4
-6
-6
-8
-8
Alternatively, the game can be written as a 22 game as follows, with strategies a1and a2 for A, and b1, b2
and b3 for B.
b1
b3
a1
-7
a2
-6
This problem is the same as in above previous example

4 (6)
x=
(8 4) (7 6)
2
=
5
258
4 (7)
(8 4) (7 6)
11
=
25
8 X 4 (7) X (6)
V=
(8 4) (7 6)
2
=5
y=
Thus, optimal strategy for A is (
2 3
11
14
, ) and for B is ( , 0 ,
,0).
5 5
5
25
11.8 Summary
In this unit the concept of Queue System and its difference with the Queue has been discussed. The assumptions
under which the models of Queue specially the single service stations (MMI/ FIFO) and multiple service
station (MMK/ FIFO) have been discussed. Their applications in practice with numerical examples have
also been explained.
In Game Theory, the concept of Pure and Mixed Strategies, the concept of Saddle Point and Principles of
Dominance have been explained .The applications of Games under competitive situations have been
discussed. The methods of solutions for 2 X 2, 2 X n, m X 2 and the L.P formulation have also been
discussed with examples.
11.9 Key Words

Queue: A group of customers waiting for service in a system is a queue.
Waiting Time: The time spent in the queue before being serviced.
Balking: Customers decides not to join the queue.
Jockeying: Leaving the first queue and joining the other.
Reneging: Joining the queue and leaving it afterwards.
Strategy : The numbers of competitive choices that are available for a player are called strategies
to that player.
Payoff matrix: it is a tabular representation of payments that should be made at the end of a game.
Zero-sum Game: Games with the algebraic sum of gains and losses of all the players equal to zero
are called zero-sum game.
n-person game: a game having n persons.
Saddle Point: is a point in the domain of a function that is a stationary point but not a local
extremum (maxima and minima).
Pure strategy: provides a complete definition of how a player will play a game. In particular, it
determines the move a player will make for any situation he or she could face. A players strategy
set is the set of pure strategies available to that player.
259
Mixed Strategy: is an assignment of a probability to each pure strategy. This allows for a player
to randomly select a pure strategy. Since probabilities are continuous, there are infinitely many
mixed strategies available to a player, even if their strategy set is finite.
Dominance: Since all players are assumed to be rational, they make choices which result in the
outcome they prefers most, given what their opponents do. In the extreme case, a player may have
two strategies A and B so that, given any combination of strategies of the other players, the outcome
resulting from A is better than the outcome resulting from B. Then Strategy: A is said to dominate
strategy B. A rational player will never choose to play a Dominated strategy.

1.
Write short note on the following

a.
Queue Discipline
b.
Queuing System
c.
Service mechanism
2.
Explain the meaning of a queue and state the object of queuing analysis. Briefly describe the elements
of queuing system.
3.
At a one-man barbershop, it takes on the average one-half hour to service a customer and the
customer arrive at an average rate of one every forty five minutes. If a Poisson exponential model is
assumed then find
4.
a.
The average time a customer spends in the shop.
b.
The average time the customer spends waiting in the queue.
c.
The expected time the barber will remain idle.
d.
At any given time, expected number of the customer in the shop.
e.
At any given time, expected number of the customer in the queue.
Given the following payoff matrix

Union strategies
X
Y
Z
W
Company strategies
C1
C2
+20
+15
+25
+14
+40
+2
-5
+4
C3
+12
+8
+19
+11
C4
+35
+10
+5
0
Determine the optimum strategies for the union as well as for the company. What is the expected
gain to the union of the above game?
5.
For what type of business problems might game theory be helpful? Explain.
6.
The matrix of a certain two-person zero-sum 3X3 game is given below

Y
260
What should be the optimal strategies for X and Y in the above game and the value of the game?
11.11 References
Business Statistics by D.R.Agarwal, VRINDA PUBLICATIONS (P) Ltd.
Business Statistics by G.C.Beri, Tata McGraw Hill Publishing Company Ltd.
Fundamentals of Statistics by S.C.Gupta, Himalaya Publishing House.
Business Statistics by S.P.Gupta, Sultan Chand & Sons Educational Publishers, New Delhi.
Statistics for Business and Economics by R.P.Hooda, MACMILLAN India Ltd.
Quantitative Techniques by C.R. Kothari, Vikas publishing house Pvt. Ltd.
261

QT Detailed 2 PDF

Uploaded by

Copyright:

Available Formats

QT Detailed 2 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

QT Detailed 2 PDF

Uploaded by

Copyright:

Available Formats

MP- 204

Vardhaman Mahaveer Open University, Kota

Vardhaman Mahaveer Open University, Kota

Vardhaman Mahaveer Open University, Kota

BLOCK - I : BASIC MATHEMATICS

Quantitative Techniques in Business

Elementary Calculus and Matrix Algebra

Correlation and Simple Regression

Business Forecasting and Time Series

BLOCK - III : PROBABILITY DISTRIBUTIONS

Probability and Probability Distributions

BLOCK - IV : OPERATIONS RESEARCH

Linear Programming and Transportation Models

Queuing and Game Theory

Unit-1 : Quantitative Techniques in Business

After reading this unit, you should be able to understand the:

Meaning of Quantitative Techniques

Classification of Quantitative Techniques

Role of Quantitative Techniques

Limitations of Quantitative Techniques

Functions and their applications

Construction and Classification of functions

Functions related to economics and some special functions

Break Even Analysis

Sequence and Series

Arithmetic Progression, Sum of A.P. Series, Arithmetic Mean

Geometric Progression, Sum of G.P. Series, Geometric Mean

Applications of A.P. and G.P.

Classification of Quantitative Techniques

(b) Programming Techniques

Methods of collecting data

Classification and tabulation of

Probability theory and Sampling

Correlation and regression

Time series Analysis

Interpolation and Extrapolation

Survey Techniques and

a. Monte Carlo Techniques

Statistical quality control

Statistical Inferences and

Waiting Line (queuing) Theory

Integrated Production Models

Network Analysis/ PERT

Role of Quantitative Techniques

Functions and Their Applications

It can be combined into a single representation by writing

f = {(-1, -1), (2, 5), (3, 7)}

3. 2. Average Cost Function

30x = 18,000x = (18,000 / 30) = 600

This is usually written as

Functions like f(x) = x3 + 2 x

Profit = Total Revenue Total cost

[Substituting q = (150 3p) given]

= -3p2 + 162p -600 which is a quadratic function in p.

4. Convex Set and Convex Function:

These functions are related to as follows:

x = g (y) = g [f (x)] or gof (x)

2 x 2 7 x 6 are examples of irrational functions.

where a, b, e, and k are constants and x is exponent.