MTH 281 MODULE 1

f (b) f (a )
F(x) = f(x) – y = f(x) – f(a) – (x – a) (36)
b a
Differentiating we have

f (b) f (a )
F(x) = f(x) – (37)
b a
which is a defined quantity in a < x < b.

Also F(a) = F(b) = 0 since the curve AB and the straight line AB
intersect at these points. Hence the function F(x) satisfies Rolle’s
Theorem and consequently there exists a value of x (say x = c)
such that F’(c) = 0. This implies (from (37)) that there exists a
value (x = c) such that

f (b) f (a )
f’(c) = (38)
b a
which proves (34).

Example 5: If f(x) = sin 3x, and a = 0, b = /6, c can be found from the
equation (see (34) or (38))

sin( / 2) sin 0
3 cos 3c = . (39)
( / 6) 0
This gives directly c = 1
3 cos-1 (2/ ).

The first Mean-Value Theorem is useful in many ways; in particular in

establishing inequalities between functions. For example, a typical
problem would be to show that in the interval 0 < x < /2

sin x
1> > 2. (40)
x

This is an extension of the inequality relation already obtained

graphically in Chapter 2. Problems like this maybe conveniently dealt
with by using the following result:

If f(x) is continuous in the range a x b, and differentiable in a < x <

b, and if f’(x) > 0 in a < x < b, then for a < x1 < x2 < b
f(a) < f(x1) < f(x2) < f(b).

Similarly if f’(x) < 0 in a < x < b, then

f(a) > f(x1) > f(x2) > f(b).
for a < x1 < x2 < b.

15
MTH 281 MATHEMATICAL METHODS 1

These statements are obvious when represented graphically (see Figs.

1.9 and 1.10), but we indicate an analytical proof here.

y
f(b)

f(x2)
f(x1)
f(a)

0 a x1 x2 b x
y = f(x), f’(x) > 0
y
f(a) Fig. 1.8

f(x1)
f(x2)
f(b)

0 a x1 x2 b x
y = f(x), f’(x) < 0
Fig. 1.9

Consider the case when f’(x) > 0. The first Mean-Value Theorem gives

f(x1) – f(a) = (x1 – a) f’(c), (41)

Where a < c < x1. But if f’(x) > 0, then f’(c) > 0. Also, by assumption, x1
> a. Hence

f(x1) > f(a) (42)

Similarly f(x2) > f(x1) and f(b) > f(x2), and hence the statement is
proved. A similar proof exists when f’(x) < 0.

Example 6: Consider now the inequality relation (40). Here

sin x
f(x) = , and f(x) 1 as x 0.
x

Differentiating we have

f’(x) = x cos x sin x , (43)

x2
which is negative in the range 0 < x < /2. Hence according to the
results obtained above we have

16
MTH 281 MODULE 1

f(0) > f(x) > f( /2), (44)

sin x
whih gives 1 > > 2. (45)
x

3.6 Higher Derivatives and Leibnitz’s Formula

When a function y = f(x) is differentiated more than once with respect to

x, the higher differential coefficients are written as

d2y d dy d 3 y d d2 y dn y d dn 1y
= , = , …, = ,
dx 2 dx dx dx 3 dx dx 2 dx n dx dx n 1

dn y
where is the nth differential coefficient of y with respect to x.
dx n
(These are sometimes abbreviated to either

f’’(x), f’’’(x) … f(n)(x)

or D2y, D3y …. Dny,

where D d/dx.)

We now give a few examples showing how the nth differential

coefficients of some simple functions maybe obtained.

Example 7: If y = sin x, then

dy
Dy = cos x = sin x ,
dx 2
d2 y
D2 y = -sin x = sin ( + x),
dx 2

d3y
Dy 3
3
= -cos x = sin 3 x ,
dx 2

and in general

n dn y
Dy = sin x . (46)
dx n 2

17
MTH 281 MATHEMATICAL METHODS 1

Example 8: If y = loge x, then

Dy = 1/x,

D2y = 1/x2,

D3y = 2/x3,

( n 1)!
and Dny = (-1)n-1 . (47)
xn

(Equation (47) is valid for all n, including n = 1, if by 0! We mean

unity.)

In these two examples, the functions have been simple enough to enable
the nth differential coefficient to be written down in a few lines. When,
however, the nth differential coefficient of a product of two functions
u(x) and v(x) is required it is better to proceed as follows:

We have shown earlier from first principles that

D(uv) = u Dv + v Du (48)

Differentiating (48) now gives

D2(uv) = u D2v + 2Du . Dv + v D2u. (49)

Similarly we obtain

D3(uv) = u D3v + 3Du . D2v + 3D2u . Dv + v D3u, (50)

D4(uv) = u D4v+4Du . D3v+6D2u .D2v+4D3u .Dv+Dv + v D4u (51)

and so on.

By inspection of these results the following formula (due to Leibnitz)

may be written down for the nth differential coefficient of uv:

Dn(uv) = u Dnv + nC1Du . Dn-1v + nC2D2u . Dn-2v + …

+ nCn-1Dn-1u.Dv + v Dnu, (52)

n!
where nCr = .
( n r )!r!

This maybe written more concisely as

18
MTH 281 MODULE 1

n
Dn(uv) = n
CrDn-rv . Dnu. (53)
r 0

Leibnitz’s formula (52) may be proved by induction as follows. Suppose

(52) is true for one value of n, say m; then by differentiating we find

Dm+1(uv) = (u Dm+1v + Du . Dmv) + mC1(Du . Dmv + D2u . Dm-1v)

+ mC2(D2u . Dm-1v + D3u . Dm-2v)+ … + (Dv . Dmu + vDm+1u), (54)

= uDm+1v+(1+mC1)Du . Dmv+(mC1+mC2)D2u . Dm-1v+…+vDm+1u. (55)

Now mCr-1 + mCr = m+1Cr (56)

and hence (55) becomes

Dm+1(uv)= uDm+1v+m+1C1Du . Dmv+ m+1C2D2u . Dm-1v+…+ vDm+1u. (57)

This again is the Leibnitz’s formula (52) with m + 1 in place of m.

Hence if the formula is true for n = m, it is certainly true for n = m + 1.

However, we know (from first principles) that it is true for n = 1, and

therefore it is true for n =2, 3, …, and consequently for all positive
integral values of n.

Example 9: To obtain the nth differential coefficient of y = (x2 + 1)e2x

we put x2 + 1 = u and e2x = v. Then by (52)

(x2 + 1)2ne2x + 2nx . 2n-1e2x + n(n – 1)2n-2e2x (58)

= 2n-2 e2x (4x2 + 4nx + n2 – n + 4). (59)

Example 10: The nth differential coefficient of y = x logex may be

obtained by putting x = u and logex. Equation (52) then gives
( n 1)! (n 2 )!
n n n-1 n-1
D y = D (uv) = x(-1) + n(-1) (60)
xn xn 1

(n 2 )!
= (-1)n-2 , (n 2). (61)
xn 1

Example 11: Leibnitz’s formula may also be applied to a differential

equation to obtain a relation between successive differential coefficients.
As this forms a step towards finding power series solutions of certain
types of differential equations now consider the following problem.
Suppose y satisfies the equation

19
MTH 281 MATHEMATICAL METHODS 1

d2y
2
+ x2y = sin x. (62)
dx

Then differentiating each term n times (using Leibnitz’s formula for the
product term x2y), we obtain (using (46))

Dn+2y + (x2Dny + 2nx Dn-1y + n(n – 1)Dn-2y) = sin n x , (63)

2

which is a relation between the (n – 2)th, (n – 1)th, nth and (n + 2)th

differential coefficients of y for all x. If we now put x = 0 in (63) we
find

0 if n is even ,
Dn+2y + n(n – 1)Dn-2y = (64)
1if n is odd .

Remark: Relations of the type of the expression in (63) and (64) at x =

0, are useful in developing power series solutions of differential
equations.

SELF ASSESSMENT EXERCISES 4

dy d2x
1) If y is a function of x, show by putting = p, that =
dx dy2
d2 y
2
- dx .
3
dy
dx

d2y dy 2
2) If y = sin ( x 1) prove that 4(x + 1) 2
+2 + y = 0.
dx dx

3.7 Maxima and Minima

A particular important application of differentiation is to the problem of

finding the maxima and minima values of a given function f(x) in same
interval a x b. Purely on geometrical grounds we can see (Fig.
3.11) that provided f(x) is differentiable in the range (a,b) then at a
maximum or minimum the tangent to the curve must be parallel to the x-
axis. According a necessary condition for a point x0 (say) to be a
maximum or a minimum is that

20
MTH 281 MODULE 1

df ( x )
f’(x0) = 0 i.e. 0at x x0 . (65)
dx

Such points are called critical points.

y
y = f(X)
Q
P

x
0 x=a x=b
Fig. 1.10

Although at the end points x = a and x=b it would seem that the function
possesses larger and smaller values respectively than at the maximum
point Q and the minimum point P, we do not count these as true maxima
and minima but note that they are just the greatest and least values of the
function in the range a x b.

Now suppose f’(x) > 0. Then the function y = f(x) increases with
increasing x. if f’’(x) > 0 then f’(x) is also increasing and hence the
curve is concave is upwards (as near the maximum point Q). Hence if
f’(x0) = 0 and f’’(x0) > 0 the point x0 is a minimum point, whilst if f’(x0)
= 0 and f’’(x0) < 0 the point x0 is a maximum point. It may happen that
both f’(x0) and f’’(x0) vanish (for example, f(x) = x3 has a critical point
at x = x0 = 0, and f’’(0) = 0). Such points are called points of inflection.
A more detailed theory based on Taylor series (see Unit4) enables the
nature of a critical point to be determined when the first n (say)
derivatives vanish at the critical point. However, we shall not deal with
this situation here.

Finally, we note that we have so far assumed that the function is

continuous and has a continuous first derivative. If the function is not
differentiable then it may still possess maxima and minima but they
cannot be found by differentiation. For example, the function y = x is
shown in Fig. 3.12. This is not a differentiable function for the range –a
x a (say). However, a true minimum does exist at x = 0.

21
MTH 281 MATHEMATICAL METHODS 1

y
y =x

x
0

Fig. 1.11

SELF ASSESSMENT EXERCISES 5

i. Determine the maxima and minima value (if any) of

a) sin-1 (x2 + 2)

b) 1 + x2/3

ii. Find the critical point of y = x2e-x and determine whether they are
maxima or minima.

4.0 CONCLUSION

In this unit, we have dealt with limit, continuity and have established
some theorems, such as Rolle’s Theorem and Mean-Valued theorem.

We have also established relationship between continuity and

differentiability.

The concept of differentiability allows us to determine, the minimum

and maximum point of a given function.

5.0 SUMMARY

Here you have learnt about limits, continuity and differentiability. You
have also learnt that differentiability at a point implies continuity at that
point.

Some relevant theorems such as Rolles and Mean-valued theorem were

also studied.

You are to master these areas in order to be able to follow the

presentation in the next unit.

22
MTH 281 MODULE 1

6.0 TUTOR-MARKED ASSIGNMENT

i. Find the derivatives of

1
a)
( ax b )

b) x4logex

ii. Find the points of discontinuity of the following functions

4x 6
a)
x2 6x 8

b) sec x

sin x
c)
x

d2y dy 2
iii. If y = sin ( x 1) prove that 4(x + 1) 2
+2 + y = 0.
dx dx

d2y dy
iv. If y = sin x = (sin x) , prove that (1 – x )
-1 -1 2 2
- x is
dx 2 dx
independent of x.

7.0 REFERENCES/FURTHER READING

G. S. Stephenson: Mathematical Methods for Science Students.

Longman, London and New York, 1977.

P. D. S. Verma; Engineering Mathematics Vikas Publishing House,

PVT Ltd., New Delhi, 1995.

23
MTH 281 MATHEMATICAL METHODS 1

UNIT 2 PARTIAL DIFFERENTIATION

CONTENTS

1.0 Introduction
2.0 Objectives
3.0 Main Content
3.1 Functions of Several Independent Variables.
3.2 First Partial Derivatives.
3.3 Function of a function.
3.3.1 Commutative Property of Partial Differentiation
3.4 Higher Partial derivatives.
3.5 Total Derivatives.
3.6 Implicit Differentiation.
3.6 Higher Total Derivatives.
3.7 Homogeneous Functions.
3.8 Euler’s Theorem.
3.9 Change of Variables.
3.10 Taylor’s Theorem for Function of two Independent
Variables.
3.11 Maxima and Minima of function of two Variables.
3.12 Lagrange Multipliers
4.0 Conclusion
5.0 Summary
6.0 Tutor-Marked Assignment
7.0 References/Further Reading

1.0 INTRODUCTION

In Unit 1, we discussed the concepts of continuity and differentiability

of one real independent variable. In this unit, we shall consider and
extend the idea developed in Unit 1 to function of more than one
variable.

This is very important because, in scientific analysis of a problem, one

often find that a factor depends upon several other factors. For example,
volume of a solid depends upon its length, breath and height. Strength of
a material depends upon temperature, density, isotropy, softness etc.

It is therefore necessary to define function of several variables. If a

variable z depends for its value upon those of x and y, we say z is a
function of x and y, and write z = f(x,y).

All types of concepts for functions of one variable are extend to

functions of several variables. For example the value of a function f(x,y)

24