QUANTITATIVE TECHNIQUES FOR Dr.

Pritha Guha
MANAGERIAL DECISION - 1 (QTMD1G21-1)
15-07-2021 1
PROBLEM 5.10, TEXTBOOK PG. 212
Trains headed to a destination A arrive at station at 15-minute intervals starting at 7 A.M.,
whereas trains headed to destination B arrive at 15-minute intervals starting at 7:05 A.M.
a) A certain passenger arrives at the station at a time that is uniformly distributed between 7 and
8 A.M, and then gets on the first train that arrives. What is the probability that the passenger
travels to A?

15-07-2021 2
PROBLEM 5.10, TEXTBOOK PG. 212
Trains headed to a destination A arrive at station at 15-minute intervals starting at 7 A.M.,
whereas trains headed to destination B arrive at 15-minute intervals starting at 7:05 A.M.
b) A certain passenger arrives at the station at a time that is uniformly distributed between 7:10
and 8:10 A.M, and then gets on the first train that arrives. What is the probability that the
passenger travels to A?

15-07-2021 3
EXPONENTIAL DISTRIBUTION
Used to model the (waiting) time between successive events, e.g., the time between
failures of light bulbs, time between two earthquakes etc.

p.d.f./density

X ~ Exp(λ), where, λ is the “rate of events”.

15-07-2021 4
 1
•Mean: E X =
λ
EXPONENTIAL •Variance: Var X =
1
𝜆2
DISTRIBUTION •c.d.f. F t = P X ≤ 𝑡 = ቊ1 − e −λt
,x ≥ 0
0, otherwise
•Survival function: S(t) = P(X > t) = 1 – F(t) = e-λt

15-07-2021 5
 •Memoryless Property:
If X ~ Exp (λ) for some λ > 0, then
P X > t + s T > t = P X > s , where s ≥ 0, t ≥ 0.
which means given survival to time t, the chance of surviving a

EXPONENTIAL further time s is the same as the chance of surviving to time s in the
first place.

DISTRIBUTION R:
p.d.f.: dexp(t, λ)
c.d.f.: pexp(t, λ)
Quantile: qexp(p, λ)

15-07-2021 6
EXAMPLES:

Banks/supermarkets - Computers - waiting for a

Failure situations - waiting

15-07-2021 7
PROBLEM 5.13, TEXTBOOK PG. 218
The lifetime of a particular mobile phone follows an exponential distribution with mean 30
months. If a person buys a second-hand mobile phone which has been used for some time, what is
the probability that it will work for an additional 12 months?

15-07-2021 8
PROBLEM 5.33, TEXTBOOK PG. 214
A toy manufacturing company sells a toy train which has a mean lifetime of 10 months. If the
lifetime of the toy train follows an exponential distribution, what should be the guarantee period
offered on the train if they do not want to replace more than 10% of the toys?

15-07-2021 9
 Counts of Arrival:
• The distribution of arrivals in a fixed interval of a particular
length is Poisson
• The number of arrivals in disjoint time intervals are
independent.
ARRIVALS AND Times Between Arrivals:
TIME BETWEEN • The distribution of waiting time until the first arrival is
exponential,
ARRIVALS • The waiting time until the first arrival and the subsequent
waiting times between each arrival and the next are
independent, all with the same exponential distribution.
Note:
Expected waiting time between two events is 1/λ.
Number of events in unit time has Poisson(λ) distribution!

7/15/2021 10
 Found almost everywhere: considered to be the ‘natural
distribution’ for a number of features for large groups: e.g.,
Height, Weight, Blood Pressure, Grades ...

NORMAL 30
25

DISTRIBUTION
20

Frequency
15
Frequency
10
5
0
Bin

Averages for large groups have approximate normal

15-07-2021 11
NORMAL DISTRIBUTION
•The pdf of normal distribution with mean μ
and variance σ2 is
−(𝑥−𝜇) 2
1
f(x)= 𝑒 2𝜎2 , for all real x
𝜎 2𝜋

•If X is normal with mean μ and variance

15-07-2021 12
NORMAL DISTRIBUTION: PROPERTIES
•Normal pdf is symmetric around its mean
μ, and its shape depends on sd σ. The
higher the sd, the flatter the curve.
•If X~N(μ, σ2), then aX+b~N(aμ+b,a2σ2).

15-07-2021 13
STANDARD NORMAL DISTRIBUTION (Z)
•If μ=0 and σ2=1, then X is the standard
normal distribution, more commonly
denoted by Z.
•We denote its pdf by φ(.); and cdf by
Φ(.).

15-07-2021 14
 For any normal random variable X, if X~N(μ,σ2), then
Z=(X-μ)/σ ~N(0,1), standard normal.

STANDARD Notations: pdf: φ(.); cdf : Φ(.)

NORMAL 1.The pdf φ(.) is symmetric around 0, i.e. φ(-z) = φ(z).

DISTRIBUTION (Z) 2.Φ(0) = 0.5 and Φ(z) + Φ(-z) = 1.

15-07-2021 15
Normal Probability Distribution
Benchmarking of normal probabilities:
99.72%
95.44%
68.26%

m
x
m – 3s m – 1s m + 1s m + 3s
m – 2s m + 2s

15-07-2021 16
15-07-2021 17
 99.72%
EXAMPLE 95.44%
68.26%
X~N(µ = 3,σ2 = 22).
a) What is P(X ≤ 3)?
b) What is P(X ≤ 7)?
c) What is P(X < 7)?
d) What is P(-3 < X ≤ 7)?
e) What is P(-3 < X ≤ 2)?
--- Need to use standard normal table now. m
x
m – 3s m – 1s m + 1s m + 3s
m – 2s m + 2s

15-07-2021 18
 Table 5.1, Page 190, Text
STANDARD NORMAL VALUES
Φ(Z)
Φ(z) is the area under the standard normal
pdf φ(.) up to z.
Notice:
a)Φ(0) = 0.5.
b)Table in book (see left) only lists Φ(z) for
z>0.
Use Φ(-z) = 1- Φ(z) for z<0.

15-07-2021 19
EXAMPLE
X~N(µ = 3,σ2 = 22).
e) What is P(-3 < X ≤ 2)?

15-07-2021 20
EXAMPLE
X~N(µ = 3,σ2 = 22).
f) What is the 95th percentile of X?
Note: 95th percentile/0.95th ‘quantile’:
x s.t. F(x) = 0.95 = F-1(0.95)

15-07-2021 21
15-07-2021 22
R:
pnorm(x,μ,s) normal probability (c.d.f.)
qnorm(x,μ,s)  normal quantile
pnorm(x) standard normal probability (c.d.f.)
qnorm(x)  standard normal quantile

15-07-2021 23
PROBLEM 5.21, TEXTBOOK PG. 213
Suppose in an apartment complex consisting of 45 families, the total drinking water usage for a
day was 1350 lit. If the water usage per family is distributed according to normal distribution
with a standard deviation 5 lit,
a) what percentage of families used more than 35 lit of water?

15-07-2021 24
PROBLEM 5.21, TEXTBOOK PG. 213
Suppose in an apartment complex consisting of 45 families, the total drinking water usage for a
day was 1350 lit. If the water usage per family is distributed according to normal distribution
with a standard deviation 5 lit,
b) what is the probability that exactly 5 families used more than 35 lit of water?

15-07-2021 25
RESULT
For two independent normal random variables X~N(µ = a, σ2 = v) and Y~N(µ = b, σ2 = u),
1. X+Y ~N(µ = a+b, σ2 = v+u).
2. X-Y ~ N(µ = a-b, σ2 = v+u).

Example: X~N(µ = 5, σ2 =16), Y~N(µ = -5, σ2 = 9), independent.

15-07-2021 26