4 Integral Equations by D.C. Sharma M.C. Goyal

Integral Equations
Integral Equations
D.C. Sharma
Head, Department of Mathematics
Dean, School of Mathematics, Statistics and Computational Sciences
Central University of Rajasthan, Kishangarh, Ajmer
M.C. Goyal
Dean (Academics), Professor and Head of Department of Mathematics
Rajasthan College of Engineering for Women, Jaipur
Delhi-110092
2017
INTEGRAL EQUATIONS
D.C. Sharma and M.C. Goyal
© 2017 by PHI Learning Private Limited, Delhi. All rights reserved. No part of this book may
be reproduced in any form, by mimeograph or any other means, without permission in writing
from the publisher.
ISBN-978-81-203-5280-3
The export rights of this book are vested solely with the publisher.
Published by Asoke K. Ghosh, PHI Learning Private Limited, Rimjhim House, 111, Patparganj
Industrial Estate, Delhi-110092 and Printed by Mudrak, 30-A, Patparganj, Delhi-110091.
Contents
Preface................................................................................................................ ix
Acknowledgements.............................................................................................. xi
1. Basic Concepts.............................................................1–11
1.1 Introduction 1
1.2 Abel’s Problem 1
1.3 Initial Value Problem and Boundary Value Problem 2
1.4 Integral Equation 3
1.5 Special Kinds of Kernels 4
1.6 Classification of Integral Equation 5
1.7 Iterated Kernels 6
1.8 Reciprocal Kernal or Resolvent Kernel 7
1.9 Eigenvalues and Eigenfunctions 7
1.10 Solution of an Integral Equation 8
Exercise 1.1 11
2. Applications to Ordinary Differential Equations....... 12–28

2.1 Introduction 12
2.2 Method of Conversion of an Initial Value Problem to a
Volterra Integral Equation 12
Exercise 2.1 17
2.3 Alternate Method of Transforming the Initial Value
Problem into a Volterra Integral Equation 18
Exercise 2.2 21
2.4 Boundary Value Problem and its Conversion to Fredholm
Integral Equation 21
Exercise 2.3 27
v
vi Contents
3. Solution of Homogeneous Fredholm Integral

Equations of the Second Kind .................................. 29–40
3.1 Introduction 29
3.2 Characteristic Value (or Eigenvalue) and Characteristic
Function (or Eigenfunction) 29
3.3 Solution of Homogeneous Fredholm Integral Equation
of the Second Kind with Separable (or Degenerate) Kernel 31
3.4 Orthogonality of Two Functions 33
3.5 Orthogonality of Eigenfunctions 33
3.6 Real Eigenvalues 34
Exercise 3.1 40
4. Fredholm Integral Equations with

Separable Kernels....................................................... 41–55
4.1 Introduction 41
4.2 Solution of Fredholm Integral Equation of the Second Kind with
Separable (or Degenerate) Kernel 41
Exercise 4.1 54
5. Integral Equations with Symmetric Kernels............... 56–83

5.1
Introduction 56
5.2
Symmetric Kernal 56
5.3
Regularity Condition 57
5.4
Inner or Scalar Product of Two Functions 58
5.5
Orthogonal System of Functions 59
5.6
Fundamental Properties of Eigenvalues and
Eigenfunctions of Symmetric Kernels 59
5.7 Hilbert–Schmidt Theorem 60
5.8 Schmidt’s Solution of Non-homogeneous Fredholm
Integral Equation of the Second Kind 63
Practice Questions with Intermediate Results 77
Exercise 5.1 83
6. Solution of Integral Equations of the Second

Kind by Successive Approximation.......................... 84–120
6.1 Introduction 84
6.2 Iterated Kernel or Function 84
6.3 Resolvent Kernel or Reciprocal Kernel 85
6.4 Solution of Fredholm Integral Equation of the Second
Kind by Successive Substitution 85
6.5 Solution of Volterra Integral Equation of the Second
Kind by Successive Substitutions 87
Contents vii
6.6 Solution of Fredholm Integral Equation of the Second
Kind by Successive Approximations: Iterative Method
(Iterative Scheme) Neumann Series 89
6.7 Resolvent Kernel of a Fredholm Integral Equation 91
6.8 Illustrations Based on the Solution of Fredholm Integral
Equation by Successive Approximations (Iterative Method) 92
Exercise 6.1 101
Exercise 6.2 103
6.9 Reciprocal Functions 104
6.10 Another Approach to Solve Fredholm Integral
Equation of the Second Kind (Volterra Solution) 105
6.11 Solution of Volterra Integral Equation of the Second
Kind by Successive Approximations: Iterative Method
(Neumann Series) 107
6.12 Resolvent Kernel and Volterra Integral Equation 109
6.13 Illustrations to Explain the Solution of Volterra Integral
Equation by Successive Approximations (or Iterative
Method) 110
Exercise 6.3 119
7. Classical Fredholm Theory..................................... 121–156

7.1 Introduction 121
7.2 Fredholm’s First Theorem 121
7.3 Working Rule for Evaluating the Resolvent Kernel and
Solution of Fredholm Integral Equation of the Second
Kind by Using Fredholm’s First Theorem 128
7.4 Fredholm’s Second Fundamental Theorem 147
7.5 Fredholm’s Third Theorem 152
Exercise 7.1 155
8. Integral Transform Methods.................................. 157–183

8.1 Introduction 157
8.2 Singular Integral Equation 158
8.3 Laplace Transform 159
8.4 Some Important Properties of Laplace Transform 159
8.5 Inverse Laplace Transform 161
8.6 Some Important Properties of Inverse Laplace Transform 161
8.7 Convolution of Two Functions 162
8.8 The Heaviside Expansion Formula 163
8.9 The Complex Inversion Formula 163
8.10 Integral Equations in Special Forms 163
viii Contents
8.11 Application of Laplace Transform to Find the Solutions of Volterra

Integral Equation 164
8.11.1 Convolution Type Kernels of Volterra Integral Equation:
Working Procedure 164
8.11.2 Resolvent Kernel of Volterra Integral Equation by Using
Laplace Transform 165
8.11.3 Solution of Integral Equations of the Type
by Using Laplace
Transform: Working Procedure 166

8.12 Fourier Transforms and Their Important Properties 177
8.13 Application of Fourier Transform to Determine the Solution of
Singular Integral Equations 179
Exercise 8.1 180
Index 185–187
Preface
Integral Equations are being used as an essential effective tool by the

mathematicians, engineers and theoretical physicists to understand and solve
research problems in their field. Two-point initial value problems and boundary
value problems with fixed and variable boundary are often encountered by
researchers.
Integral equations have now been established as a far effective and highly
useful branch in almost all disciplines of knowledge. Hence, this course has
been taken as an indispensable part of syllabi of all Indian universities at
postgraduate level. Not only that, in competitive examinations like NET/SET,
this course inherits importance. Therefore, it becomes necessary for students
and teacher alike to follow the concepts and operations of integral equations.
Looking to these requirements, we have tried to provide the concepts and
principles quite clearly and in an organised manner. In order to make the book
user-friendly, we have presented the matter in a simple way and it is complete
in all respects from the examination point-of-view. By presenting sufficient
number of assorted examples, we have tried to inculcate the habit and create
confidence in students to try to solve more and more problems on their own.
Because of author’s long experience of teaching and also being actively
engaged in research, it is expected that the book shall prove immensely
beneficial to the students for whom it is meant.
All suggestions for the improvement of the book shall be thankfully
acknowledged.
D.C. Sharma
M.C. Goyal
ix
Acknowledgements
We wish to record our sincere thanks to the authors of many earlier standard
works, which have been consulted by us during the preparation of the book.
Our heartfelt thanks are due to Prof. G.C. Sharma and Prof. K.C. Gupta,
whose inspiration prompt us to take this project of writing on this topic.
We express our personal and sincere thanks to our senior colleagues and
students for their useful suggestions in preparing the blueprint and subject
matter of the book. We further reiterate that without fruitful discussion with
them, it would have been impossible to present this book.
We are also thankful to our family members for taking all the care during
long hours at the desk.
Finally, the authors are especially thankful to PHI Learning for its
cooperation in all aspects of the production of this book.
D.C. Sharma
M.C. Goyal
xi
Chapter 1
Basic Concepts
1.1 INTRODUCTION
Various physical problems in physics and other applied fileds culminate into
initial value problems or boundary value problems. Although it is equivalent to
frame the problem in the form of (ordinary and partial) differential equations
or in the form of integral equations, but it is preferred to choose the integral
form due to two main reasons. Firstly, the solution of integral equation is much
easier than the original boundary value or the initial value problems. The second
reason lies in the fact that integral equations are better suited to approximate
methods than differential equations. Moreover, integral equations develop as
representation formulae for the solution of differential equations. We will find
in the forthcoming chapters that differential equations can be replaced by an
integral equation with the help of initial and boundary conditions. As a result,
each solution of the integral equation satisfies the boundary conditions itself.
1.2 ABEL’S PROBLEM

In 1826, Abel obtained an integral equation by considering the motion of a
material point P(x, y) under the action of force of gravity moving in vertical
plane (x, h) along some smooth curve. It is required to establish the curve
such that the material point P, starting from rest at P(x, y) reaches the point
Q(x, h) at any instant t. Let T be the time taken by the particle from P to the
lowest point O, the origin of coordinates and axes, as shown in Fig. 1.1. Let
 s, then the velocity of particle at Q is
OQ
ds
 – 2 g( x – x )
dt
Q ds
∴ [t ]T0  – 
P 2 g( x – x )
1
2 Integral Equations
P ds
∴ T 
O 2 g( x – x )
Fig. 1.1 Particle falling under gravity along the curve PQO.
Now, if the shape of curve is given, then s, and hence, ds can be expressed
in terms of x. We take
ds = u(x)dx, and then
x u(x ) dx
T 
0 2 g( x  x )

Here, if the curve is such that the time of descent T is a function of x,
say, f(x), then the above relation shapes as
x 1
f ( x)   u(x ) dx
0 2 g( x – x )

This leads to find the unknown function u(x) and we get Abel’s integral
equation.
During the analysis, it is found that an initial value problem is always
converted to a Volterra integral equation, while a boundary value problem
is always converted to a Fredholm integral equation. Before we take the
classification of integral equations, it is worth to be familiar with the initial
value problems and the boundary value problems.
1.3 INITIAL VALUE PROBLEM AND BOUNDARY VALUE PROBLEM
Initial value problem

When an ordinary differential equation is solved under conditions which involve
dependent variable and its derivative at the same value of its independent
Basic Concepts 3
variable, then the problem under consideration is said to be an initial value
problem.
For instance
y"' + xy" + 2(x2 – x)y = ex – x,
with the conditions y(0) = 1, y'(0) = –1, y"(0) = k
Boundary value problem
When an ordinary differential equation is solved under conditions which
involve dependent variable and its derivative at two different values of
independent variable, then the problem under consideration is said to be a
boundary value problem.
For example
y" + xy' + 2y = e–x
with the conditions y(0) = 0, y'(1) = –1
1.4 INTEGRAL EQUATION

An integral equation is an equation in which one unknown function (which is
to be determined) appears under one or more integral signs. If in the equation,
the derivatives of this unknown function are also present, it is called an integro-
differential equation.
For example, for a  x  b, a  x  b, the equations
b
f ( x )  l  K ( x, x )  u(x ) dx, [1.1(a)]
u( x ) 
a
b

u( x ) a K ( x, x )  u(x ) dx [1.1(b)]
b
u( x ) l  K ( x, x ) [u(x )]3 dx [1.1(c)]
a
are integral equations. u(x) is an unknown function, f(x), K(x, x) are known
functions and l, a, b are constants. The functions involved may be complex
valued functions of real variables x and x.
du b
 l  K ( x, x )  u(x
) dx f ( x ), a  x  b [1.1(d)]
dx a
is integro-differential equation of unknown function u(x).

The known bivariate function K(x, x) which is integrable in the domain
a  x  b , a  x  b is called the kernel of the equation; the known function
f(x) is a continuous function. In the book, we shall find various forms of this
kernel and will establish the solution of the integral equation consequentely.
An integral equation is called linear if only linear operations are performed
in it upon the unknown functions, which means it is the equation in which no
non-linear functions of the unknown function are involved. Equations [1.1(a)]
and [1.1(b)] are linear equations. An equation which is not linear is known as
non-linear integral equation, as shown in Eq. [1.1(c)].
The most general linear integral equation may be shaped as
v( x )  u( x )  f ( x )  l  K ( x, x )  u(x ) dx (1.2)
a
where the upper limit of the integral may be variable or constant. The functions
v, f, K are known functions and u is to be determined, l is a non-zero
parameter, which may be real or complex.
The importance of keeping l separate from K lies in the fact that it plays
an essential role in the theoretical arguments for the problem under context.
Kinds of linear integral equations
1. First kind, if v = 0, Eq. (1.2) reduces to
f ( x )  l  K ( x, x )  u(x ) dx 
0
a
2. Second kind, if v = 1, Eq. (1.2) provides
f ( x )  l  K ( x, x )  u(x ) dx
u( x ) 
a
3. Third kind, if v  0 , Eq. (1.2) itself works.
1.5 SPECIAL KINDS OF KERNELS

As mentioned earlier, the role of the known bivariate function K(x, x) is quite
significant both from the problem and its solution point-of-view. Mainly, we
shall come across with the following forms:
1. Symmetric kernel: The kernel K(x, x) is symmetric (complex symmetric
is also called Hermitian) if
K ( x, x )  K (x , x )
where the bar represents the complex conjugate. A real kernel is symmetric
x
if K(x, x) = K(x, x). For instance exx is symmetric, while tan 1   is not a
 x
symmetric kernel.
n
2. Separable or degenerate kernel: If K ( x, x )   gi ( x ). hi (x ) means K
i 1
has been expressed as the sum of a finite number of terms, each of which is
the product of function of x only and x only, then such a kernel is called
separable or degenerate kernel. Obviously, gi(x) and hi(x) are linearly
independent (or else some of the terms will combine, and consequently, the
number of terms will reduce). A degenerate kernel has a finite number of
characteristic values.
3. Difference kernel: A kernel of the form K(x – x) is called difference
kernel.
Basic Concepts 5
1.6 CLASSIFICATION OF INTEGRAL EQUATION

Integral equations are classified into the following four classes:
1. Fredholm integral equation
2. Volterra integral equation
3. Singular integral equation
4. Convolution integral equation
Fredholm integral equation
A linear integral equation of the form
b
v( x )  u( x )  f ( x )  l  K ( x, x )  u(x ) dx (1.3)
a
where a and b both are constants is called Fredholm integral equation of the
third kind. Here, f(x), v(x) and K(x, x) are known functions, u(x) is unknown
function and l is real or complex parameter.
Now, if in Eq. (1.3), we set v(x) = 0, then we get
b
f ( x )  l  K ( x, x )  u(x ) dx 
0 (1.4)
a
and it is Fredholm integral equation of the first kind.

Next, if in Eq. (1.3), v(x) = 1, we get
b
f ( x )  l  K ( x, x )  u(x ) dx (1.5)
u( x ) 
a
and it is Fredholm integral equation of the second kind. If in Eq. (1.5),
f(x) = 0, i.e.,
b
u( x ) l  K ( x, x )  u(x ) dx (1.6)
a
it is known as homogeneous Fredholm integral equation of the second kind.

Volterra integral equation
A linear integral equation of the form
x
v( x )  u( x )  f ( x )  l  K ( x, x )  u(x ) dx (1.7)
a
where the upper limit of the integral is variable, v(x), f(x) and K(x, x) are
known functions and u(x) is unknown function, is said to be Volterra integral
equation of the third kind. As usual l is a real or complex parameter and the
function K(x, x) is the kernel of the integral equation.
If we set v(x) = 0, i.e., Eq. (1.7) takes the form
x
f ( x )  l  K ( x, x )  u(x ) dx 
0 (1.8)
a
then it is called Volterra integral equation of the first kind.
Again, if u(v) = 1, Eq. (1.7) takes the shape
x
( x ) f ( x )  l  K ( x, x ) . u(x )dx (1.9)
u
a
then it is called Volterra integral equation of the second kind.
Further, if in Eq. (1.9), f(x) = 0, then it becomes
x
u( x )  l  K ( x, x ) . u(x )dx (1.10)
a
and is called homogeneous Volterra integral equation of the second kind.
Singular integral equation
An integral equation is a singular integral equation, if either
1. One or both the limits of integration are infinite, or
2. The kernal becomes infinite at one or more points within the range of
integration. For instance,
  x x
( x) f ( x)  l 
u e u(x )dx

x 1
and u( x ) = ∫ ⋅ u(ξ )dξ, 0  a  1
0 ( x − ξ )α
are both singular integral equations.
Convolution integral equations
If the kernel K(x, x) is a function of one variable and is of the type K(x, x) =
K(x – x), then the integral equations, i.e., Eqs. (1.5) and (1.9) take the shape
b
f ( x )  l  K ( x  x ) . u(x )dx (1.11)
u( x ) 
a
x
and f ( x )  l  K ( x  x ) . u(x )dx (1.12)
u( x ) 
a
respectively, and are called integral equations of convolution type.

Further, if u1(x) and u2(x) are two continuous functions defined for x  0,
then the convolution or faltung of u1 and u2 is expressed or defined by
x x
u1 
u2 o u1 ( x  x ) . u2 (x )
dx  u1 (x ) . u2 ( x  x )dx (1.13)
o
1.7 ITERATED KERNELS

1. For the Fredholm integral equation of the second kind [Eq. (1.5)], we define
the iterated kernel, Kn(x, x), n = 1, 2, 3, ... as below:
K1(x, x) = K(x, x)
b
and K n ( x, x ) 
a
K ( x, z )K n 1 ( z, x )dz, n 2,3 (1.14)
2. For the Volterra integral equation of the second kind [Eq. (1.9)], we define
the iterated kernels Kn(x, x), n = 1, 2, 3, ... as below:
K1(x, x) = K(x, x)
Basic Concepts 7
x

and K n ( x, x ) 
x
K ( x, z )K n 1 ( z, x )dz, n 2,3,… (1.15)
1.8 RECIPROCAL KERNAL OR RESOLVENT KERNEL

Let the solutions of integral equations, i.e., Eqs. (1.5) and (1.9) of the second
kind be:
b
( x ) f ( x )  l  R( x, x; l ) f (x )dx (1.16)
u
a
x
and ( x ) f ( x )  l  S ( x, x; l ) f (x )dx (1.17)
u
a
respectively, then R(x, x; l) and S(x, x; l) are called the reciprocal or resolvent
kernel of the respective integral equation.
1.9 EIGENVALUES AND EIGENFUNCTIONS

We consider the homogeneous Fredholm integral equation
b
u( x )  l  K ( x, x )u(x )dx (1.18)
a
for which u(x) = 0 is an obvious solution. This solution is taken as zero or

trivial solution.
The values of the parameter l for which Eq. (1.18) possesses non-
zero solutions [u(x) ≠ 0] are the eigenvalues of Eq. (1.18) or of the kernal
K(x, x). Further, corresponding to such eigenvalues of l, every non-zero
solution of Eq. (1.18) is called an eigenfunction.
Remarks: 1. Eigenvalues are also termed as characteristic values or characteric
numbers. Similarly, eigenfunctions are also called characteristic
functions or fundamental functions.
2. l = 0 is not considered as eigenvalue, since it corresponds to
u(x) = 0.
3. Corresponding to eigenvalue l, u(x) and C u(x) are eigenfunctions
for arbitrary constant C.
4. The set of all characteristic numbers of an integral equation with
a kernel K(x, x) is called the spectrum of the kernel (or of the
integral equation).
5. A homogeneous Fredholm integral equation may not possess any
real eigenvalue or eigenfunction.
1.10 SOLUTION OF AN INTEGRAL EQUATION

A solution of an integral equation [such as Eqs. (1.3) and (1.7)] is a function
u(x), which when substituted in the equation, reduces it to an identity. For
example, for the integral equation.
x
u( x ) 1   u(x )dx
0
the solution is u(x) = ex, which is verified immediately.

When we take the questions for different kernels, we need two specific
formulae of integration, which are as follow:
1. Formula for converting a multiple integral of order n into a single
ordinary integral of order one: It is given below:
x x ( x  x ) n 1
u(x )  dx   
n
a a (n  1)!
u(x )dx (1.19)
n n
[(dx) is also written as dx ]
2. Leibnitz’s rule of differentiation under the sign of integration: If F(x, x)
F
and are continuous functions of both x and x and if the first derivatives
x
of G(x) and H(x) are continuous, then
d H ( x) H ( x ) F dH dG

dx G ( x ) 
F ( x, x ) dx 
G ( x ) x
dx  F[ x, H ( x )]
dx
 F[ x, G( x )]
dx
(1.20)
If G(x) and H(x) are constant functions, then Eq. (1.20) reduces to
d H H F
dx G
F ( x, x )dx  
G
dx
x (1.21)
 2
EXAMPLE 1.1: Show that the function u( x ) e x  2 x   is a solution of
1
 3
Fredholm integral equation u( x )  2  e x x u(x ) dx  2 xe x .
0
 2
x
Solution: We have u( x ) e  2 x  
 3
 2
 u(x ) ex  2x   ,
 3
Then, the L.H.S. of the given integral equation
 2 1  2
 e x  2 x    2  e x x  ex  2x   dx
 3  0  3

1
 2  2x 
 e  2 x    2 e x x 2  
x
 3   3 0

 2  2
 e x  2 x    2e x 1   2 xe
x
 3  3 

Basic Concepts 9
\ L.H.S. = R.H.S.
Hence proved.
EXAMPLE 1.2: Show that the function u(x) = xex is a solution of the
Volterra integral equation.
x
sin x  2  cos( x  x ) . u(x )dx
u( x ) 
0

Solution: We have u(x) = xex
\ u(x ) x  ex
x
Then, the R.H.S. of given integral equation sin x  2  cos( x  x )  x  ex dx
x 0
 sin x  2  x  ex {cos(x  x )}dx
0
(Now, integrating by parts)

x
 ex 
 sin x  2  x  {cos(x  x )  sin(x  x )}
 1  1 0
x ex
2   1  {cos(x  x )  sin(x  x )} dx
0 11
x x
 sin x  xe x   ex cos(x  x )dx   ex sin(x  x )dx
0 0
x
 eξ x 
= sin x + xe −  {cos(ξ − x ) + sin(ξ − x )}
 2 0
x
e ξ 
−  {sin(ξ − x ) − cos(ξ − x )}
2 0
 ex 1   ex 1 
 sin x  xe x    (cos x  sin x )    ( 1)  (  sin x  cos x ) 
 2 2   2 2 
= xex = u(x)
\ L.H.S. = R.H.S.
Hence proved.
EXAMPLE 1.3: Show that u(x) = cos2x is a solution of the integral equation
x
cos x 3 K ( x, x )  u(x ) dx (i)
u( x ) 
0
sin x.cos x , 0  x  x
where, K ( x, x )   (ii)
cos x  sin x , x  x  p
Solution: We have u(x) = cos2x, which means u(x) = cos2x, and thus, the
R.H.S. of Eq. (i) shapes as [we use kernel as defined by Eq. (ii)]
xx p
 cos x  3 cos x  sin x  cos 2x dx  3 sin x  cos x  cos 2x dx
x 0x x

x p
cos x  3cos x  (cos 2x  sin x ) dx  3sin x  (cos 2x  cos x ) dx

0 x
3 x 3 p
cos x  cos x  (sin 3x  sin x ) dx  sin x   (cos3x  cos x ) dx
2 0 2 x
x p
3  1  3 1 
cos x  cos x   cos3x  cos x   sin x  sin 3x  sin x 
2  3  0 2  3 x

3  1 1 
 cos x  cos x   cos3 x  cos x   1
2  3 3 
3 1 1 
 sin x  sin 3p  sin p  sin 3 x  sin x 
2 3 3 
1 3
 cos x  (cos3 x  cos x  sin 3 x  sin x )  (cos2 x  sin 2 x )  cos x
2 2
1 3
= − cos 2 x + cos 2 x = cos 2 x = u( x )
2 2
Hence, u(x) = cos2x is a solution of the given integral equation.
px
EXAMPLE 1.4: Show that the function u( x )  sin   is a solution of the
 2 
p2 1 x
Fredholm integral equation u( x ) 
4  0 K ( x, x )u(x ) dx  2
, where the kernel
is of the form
 x (2  x )/2, 0  x  x ,
K ( x, x )  
x (2  x )/2, x  x  1
px  px 
Solution: Since u( x )  sin   , we have u(x )  sin  
 2   2
L.H.S. of the integral equation
px p2  x
K ( x, x )  u(x ) dx   K ( x, x )  u(x ) dx 
1
 sin
2
  
4  0 x 
Now, substituting for kernel K(x, x) and u(x), we get
p x p 2  x 1  px 11  px 
 sin
2
   
4  0 2
x (2  x ) sin dx    x (2  x ) sin dx 
 2 x  2  2 

px p2 x px p2 1 px
 sin  (2  x )  x sin dx  x  (2  x )sin dx
2 8 0 2 8 x 2
px p2    cos(px /2)  x x  cos(px /2)  

 sin  
(2  x ) x     0 1     dx 
2 8    p /2 0  p /2  

Basic Concepts 11
p 2 x    cos(px /2)  
1
 cos(px /2)  1
 (2  x )  
   x ( 1)   dx 
8   p /2  x  p /2  

p x p 2 (2  x )  2 x p x 2 2  px  
x
 sin   cos   sin  
2 8  p 2 p p 2 0 

p2x  2 p x 2 2  px  
1
  (2  x ) cos   sin  
8  p 2 p p 2  x 

p x p 2 (2  x )  2 x px 4 px

sin    cos  2 sin
2 8  p 2 p 2 
p2x  2 px 4 4 px
  (2  x ) cos  2  1  2 sin
8 p 2 p p 2 
 p x   (2  x ) x  x x
  sin  1   

 2  2 2  2 2
x
Thus, u( x )  is a solution of the given integral equation.
2
EXERCISE 1.1
Verify that the given functions are solutions of the corresponding integral
equations:
1 x x
(1  x 2 ) 3/2 , u( x ) 
1. u( x )  2
 u(x ) dx
1 x 0 1  x2
1
1, u( x )   x (e xx  1)u(x ) dx 
2. u( x )  ex  x
0
1 x u(x )

3. u( x )  , dx 1
p x 0 x x
1

4. u( x ) x , u( x )   K ( x, x ) u(x ) dx
0
 x (2  x )
 2 , 0 x x
x
x (4 x 3/2  7), K ( x, x ) 
15  x (2  x ) , x  x  1
 2
x3 x
5. g( x ) x  ; g( x )  x  0 sinh( x  t )  g(t ) dt
6

Chapter 2
Applications to Ordinary
Differential Equations
2.1 INTRODUCTION
The quest for establishing a representation formula to replace an ordinary
differential equation (with initial value problem or boundary value problem)
always leads to an integral equation. More specifically, it is found that an initial
value problem is converted to a Volterra integral equation, while a boundary
value problem is converted to a Fredholm integral equation. Finally, it eases
our work of finding the solution of integral equation thus obtained.
Recalling once more, in an initial value problem, the boundary conditions
are for the same value of independent variable, while in the case of boundary
value problem, the boundary conditions are for different values of independent
variable.
2.2 METHOD OF CONVERSION OF AN INITIAL VALUE

PROBLEM TO A VOLTERRA INTEGRAL EQUATION
Let an ordinary differential equation of order n be
dny d n 1 y d n 2 y
 a1 ( x )  a 2 ( x)    a n ( x)  y φ ( x ) (2.1)
dx n dx n 1 dx n  2
with the initial conditions
( n 1)

y(a) q
0 , y '( a) q1 , … , y (a) qn 1 (2.2)
where a1(x), a2(x), ..., an(x), f(x) are defined and are continuous in a  x  b.
Let u(x) be an unknown function such that
dny
 u( x )
dx n (2.3)
Integrating Eq. (2.3) from a to x, we get
12
Applications to Ordinary Differential Equations 13

x
 d n 1 y  x
 n 1   a u( x ) dx
 dx  a

d n 1 y x
or
dx n 1  u( x) dx
 y( n 1) (a) 
a
d n 1 y x
or 
dx n 1
a u( x) dx  qn 1 [using Eq. (2.2)] (2.4)
d n 1 y x
or 
dx n 1 a
u(t ) dt  qn 1 (2.5)
Integrating Eq. (2.4) with respect to x from a to x, we obtain

x
 d n 2 y  x x
a u( x)(dx)  a
2

 n 2  qn 1 dx
 dx  a
n 2
d y x
 y( n  2) a u( x)(dx)
2
or n 2
(a)  qn 1 ( x  a)
dx
d n 2 y x
a u( x)(dx)
2
or   ( x  a)qn 1  qn  2 [using Eq. (2.2)] (2.6)
dx n  2
d n 2 y x
or 
dx n 2 a
u(t )(dt )2  ( x  a)qn 1  qn  2
Now, using Eq. (1.19) (i) for the double integral, we get
d n 2 y x

dx n  2
 a ( x  t ) u(t ) dt  ( x  a) qn 1  qn 2 (2.7)
Now, integrating Eq. (2.6) with respect to x from a to x
x
 d n 3 y  x x x
a u( x)(dx)  qn 1  ( x  a)dx  qn  2  dx
3
 n 3 
 dx  a
a a

x
d n 3 y x  ( x  a) 2 
a
( n  3) 3 x
or  y ( a) u ( x )( dx )  q n 1    qn  2 [ x ]a
dx n 3  2  a
d n 3 y x ( x  a)2
a
3
or  u ( x )( dx )  q n 1  qn  2 ( x  a)  qn 3 (2.8)
dx n 3 2!
d n 3 y x ( x  a)2
a
3
or  u(t )( dt )  q n 1  ( x  a)qn  2  qn 3
dx n 3 2!
Now, applying Eq. (1.19) for the triple integral, we get
d n 3 y x ( x  t )2 ( x  a)2 ( x  a)

dx n 3
a 2!
u(t )dt  qn 1
2!
 qn  2
1!
 qn  3 (2.9)
Proceeding in this way, we get it reduced to

dy x ( x  t )n 2 ( x  a) n  2 ( x  a) n 3
  u(t )dt  qn 1  qn  2 + + q2 ( x − a ) + q1
dx a ( n  2)! (n  2)! (n  3)!
(2.10)
x ( x  t) n 1 n 1 n  2
( x  a) ( x  a)
and y  u(t )dt  qn 1  qn  2
a ( n  1)! (n  1)! (n  2)!
+ + q1 ( x − a ) + q0 (2.11)

Now, multiplying Eqs. (2.3), (2.5), (2.7), (2.9), ..., (2.11) by 1, a1(x), a2(x),
..., an–1(x) and an(x), respectively, and adding, we get
dny d n 1 y d n 2 y dy
n
 a1 ( x ) n 1
 a 2 ( x) n 2
  a n 1 ( x )  a n ( x)  y
dx dx dx dx
 u( x )  [qn 1a1 ( x )  qn  2  ( x  a)qn 1 a 2 ( x ) 
 ( x  a) n 1 
 q0  q1 ( x  a)   qn 1  a n ( x )]
 (n  1)! 
x ( x  t )2 ( x  t )n 1 
  a1 ( x )  ( x  t )a 2 ( x )  a3 ( x)   a n ( x )  u(t )dt (2.12)
a
 2! (n  1)! 
Now for L.H.S., we use Eq. (2.1) and let
y ( x ) qn 1a1 ( x )  qn  2  ( x  a)qn  2 a 2 ( x ) 

 ( x  a) n 1 
 q0  q1 ( x  a)   qn 1 a n ( x)
 (n  1)! 

 ( x  t )n 1 
and K ( x, t )  a1 ( x )  ( x  t )a 2 ( x )   a n ( x)
 (n  1)! 
so that Eq. (2.12) reduces to
x
φ ( x ) u( x )  y ( x )   K ( x, t )  u(t )dt (2.13)
a
Again, let φ ( x )  y ( x ) 
f ( x ) , then
x
f ( x )   K ( x, t )  u(t )dt (2.14)
u( x ) 
a
which is Volterra integral equation of the second kind. Thus, we have found a
relation between a linear differential equation (Eq. 2.1) and a Volterra integral
equation, which establishes that an initial value problem is converted to a
Volterra integral equation.
EXAMPLE 2.1: Transform the initial value problem

d2 y
 xy
 1, y(0) 0, y '(0)
 0
dx 2
to a Volterra integral equation of the second kind.
Solution: The given differential equation is
d2 y
 x  y  (i)
1
dx 2
subject to initial conditions
y(0) = 0, [ii(a)]
y¢(0) = 0 [ii(b)]
2
d y
Let  u( x ) (iii)
dx 2
Now, integrating Eq. (iii) with respect to x from 0 to x, we get
x
 dy  x dy x

 0

 dx   0 u( x )dx or dx  y (0) 0
u( x )dx
which on using Eq. [ii(b)], becomes

dy x
  u( x )dx (iv)
dx 0
dy x
or   u(t )dt (v)
dx 0
Now, integrating Eq. (iv) with respect to x and using Eq. [ii(a)], we obtain
x x
0 u( x)(dx) 0 u(t )(dt )
2 2
y( x ) 
 y(0) 
y( x )

which on applying Eq. (1.19) gives
x
y
( x) 0 ( x  t )u(t )dt (vi)
d2 y
Finally, putting from Eq. (iii) and y from Eq. (vi) into Eq. (i), we get
dx 2
x
u( x )  x   ( x  t )u(t )dt  
x
1 1   ( x  t )u(t )dt
or u( x ) 
 0  0
which is the required equation.
EXAMPLE 2.2: Transform the initial value problem

d3y
 2 xy 
0
dx 3
subject to initial conditions
1
 y(0) , y (0) 1,y (0) 1
2
into a non-homogeneous Volterra integral equation of the second kind.
Solution:
The given equation is
d3y
 2 xy 
0 (i)
dx 3
d3y
Let  u( x ) (ii)
dx 3
Integrating Eq. (ii) with respect to x from 0 to x, we get
x
 d2 y  x x

 dx 0

 2   0 u( x ) dx or y ( x )  y (0) 0
u( x ) dx
x
or y ( x ) 1   u( x ) dx (iii)
0
Again, integrating Eq. (iii) with respect to x from 0 to x, we obtain

x x
 y (
x )0  dx   u( x )(dx )2
x
0 0

x
or y ( x ) 1  x   u( x )(dx )2 (iv)
0
Integrating once more with respect to x from 0 to x, we get
x x x
0 dx  0 x dx  0 u( x)(dx)
3
y( x )  y(0) 

1 x2 x
or y( x ) = + x + + ∫ u(t )( dt )3
2 2 0
1 x2  x ( x  t )2
or y( x )    x     u(t )dt [using Eq. (1.19) (v)
2 2  0 2!
Now, putting the values from Eqs. (ii) and (v) in Eq. (i), we get
1 x2  x ( x  t )2
u( x )  2 x   x    2 x  u(t )dt 0
2 2 0 2!

x
or u( x )  x ( x  1)2   x ( x  t )2 u(t )dt, which is the required equation.
0
Note: In this example, we have not changed the variable of integration from
x to t for y¢¢(x) and y¢(x), since this change was not required.
EXAMPLE 2.3: Form an integral equation corresponding to the differential

equation
d2 y dy
2
 sin x  ex  y x
dx dx
with the initial conditions
y(0) = 1, y¢(0) = – 1
Solution:
d2 y dy
2
 sin x  ex  y x (i)
dx dx
Let
d2 y
 u( x )
dx 2 (ii)
Integrating Eq. (ii) with respect to x from 0 to x
x
 dy  x dy x

 0

 dx   0 u( x )dx or dx  y (0) 0
u( x )dx
dy x
or 1   u( x )dx (iii)
dx 0
dy x
or 1   u(t )dt (iv)
dx 0
Integrating Eq. (iii) again with respect to x from 0 to x, we get

x
 y( x)0x 
[ x ]0x   u( x )(dx )2
0

x
or y( x )  y(0)  x   u( x )(dx )2
0
x
or y( x ) 1  x   u(t )(dt )2
0
x
or y( x ) 1  x   ( x  t )u(t )dt (v)
0
dy d2 y
Putting the values of y, and from Eqs. (v), (iv) and (ii), respectively,
dx dx 2 x
in Eq. (i), we get u( x ) x  sin x  e x (1  x )   [sin x  e x ( x  t )]u(t ) dt , which
0
is non-homogeneous Volterra integral equation of second kind.
EXERCISE 2.1
1. Reduce the following initial value problems to Volterra integral equations

of the second kind.
d2 y dy
(a)  2 x 0, y(0) = 1, y¢(0) = 0
 3y 
2 dx
dx
y   xy   y 
(b) 0 ,
y(0) 1,
y (0) 0
y ′′′ + xy ′′ + ( x 2 − x ) y = xe x + 1 , y(0) = 1, y¢(0) = 1, y¢¢(0) = 0

(c)
d2 y dy
(d) 2  a1 ( x )  a 2 ( x ) y  F ( x ), y(0)  c0 , y (0)  c1
dx dx
Answers:
x
u( x ) 3   (5 x  3t )  u(t )dt
(a)
0
x
u( x ) 1   (2 x  t )  u(t )dt
(b)
0
x 1 
) xe x  1  x ( x 2  1)    x  ( x 2  x )( x  t )2  u(t )dt
u( x 
(c) 0  2 
x
u( x )  F ( x )  c1a1 ( x )  (c0  c1 x )a 2 ( x )   a1 ( x )  a 2 ( x ).( x  t ) u(t )dt
(d) 0
2.3 ALTERNATE METHOD OF TRANSFORMING THE INITIAL

VALUE PROBLEM INTO A VOLTERRA INTEGRAL EQUATION
We can convert a linear differential equation describing an initial value
problem into a Volterra integral equation by a different procedure also. Another
advantage with this method is that we can derive the original differential
equation with its initial conditions from the transformed integral equation. The
procedure will be clear through the examples given here.
EXAMPLE 2.4: Transform the following initial value problem into an

integral equation:
d2 y dy
2
 a ( x)  b ( x ) y g ( x ), y(a) a0 , y (a) b0
dx dx
d2 y dy
2
 g ( x )  b ( x )  y( x )  a ( x ) (i)
dx dx
and the given initial conditions are
y(a) = a0 [ii(a)]
y¢(a) = b0 [ii(b)]
Integrating with respect to x from a to x
x
 dy 
g ( x)  b ( x) y( x) dx  a a ( x)  dx
x x dy
 a

 dx  a  dx 

Now, integrating by parts,
dy x x
a g ( x)  b ( x) y( x) dx  a ( x)  y( x)a  a a '( x)  y( x) dx
x
or b0
dx
Now, using Eqs. [ii(a)] and [ii(b)],
dy x
or  b0  a ( x ) y( x )  a (a)  a0   g ( x )  b ( x ) y( x )  a ( x ) y( x ) dx
dx a
Integrating again with respect to x from a to x,

x x
[ y]ax  a b0  a (a)  a0  dx  a a ( x)  y( x) dx
x
 g ( x )  b ( x ) y( x )  a '( x )  y( x ) (dx )2
a
x
or y( x )  y(a)  {b0  a (a)  a0 }( x  a)   a (t )  y(t ) dt
a
x
  {g (t )  b (t )  y(t )  a '(t )  y(t )}(dt )2
a
Now, using Eq. (1.19), we get
x
y( x )  a0  ( x  a){b0  a (a)  a0 }   a (t )  y(t ) dt
a
x
  ( x  t ){g (t )  b (t ) y(t )  a '(t ) y(t )}(dt )
a
x
or y( x )  a0  ( x  a){b0  a0  a (a)}   ( x  t )g (t ) dt
0
x
  [a (t )  ( x  t ){b (t )  a '(t )}] y(t ) dt
a
This is the required integral equation.

x
EXAMPLE 2.5: Convert y   sin x  y   e  y  x with initial conditions
y(0)  1, y (0)  1 to a Volterra equation of the second kind. Also, derive
the original differential equation with the initial conditions from the integral
equation obtained in the previous part.
Solution: The given differential equation with the initial conditions is
y ( x ) x  e x  y( x )  sin x  y '( x ) (i)
with y(0)  1, [ii(a)]
y¢(0) = –1 [ii(b)]
Integrating Eq. (i) with respect to x from 0 to x, we get
x2 x x
 0 2 0 e x y( x) dx  0 sin x  y ( x) dx
y  ( x )
x


x2 x x
or y ( x )  1   e x  y( x ) dx  [sin x  y( x )]0x   cos x  y( x ) dx
2 0 0
x2 x
or y ( x )  1  sin x  y( x )   (e x  cos x )  y( x ) dx (iii)
2 0
Integrating Eq. (iii) with respect to x from 0 to x, we get

x
 x3  x x
[ y( x )]0x    x    sin t y(t ) dt   (et  cos t ) y(t )(dt )2
 6 0 0 0
x3 x
or y( x )  1  x   [sin t  ( x  t )(et  cos t )]y(t ) dt
6 0
 x3  x
or )   x  1   [sin t  ( x  t )(et  cos t )]y(t ) dt (iv)
y( x 
 6  0
which is a Volterra integral equation of the second kind.
Now, we differentiate Eq. (iv) with respect to x and obtain
x2 d x

y ( x )
2
1
dx 0
{sin t  ( x  t )(et  cos t )}y(t ) dt
For this, we apply Leibnitz’s rule of differentiation under the sign of

integration Eq. (1.20), and get
x2 x 
y ( x ) 1  [{sin t  ( x  t )(et  cos t )}y(t )] dt
2 0 x

d
 [sin x  ( x  x )(e x  cos x )] y( x ) ( x )
dx
d
 [sin 0  ( x  0)(e0  cos 0)] y(0)  (0)
dx
x2 x
or  1   [et  cos t ] y(t ) dt  sin x  y( x ) (v)
y ( x )
2 0
Differentiating Eq. (v) with respect to x

d x t
dx 0
y ( x ) 
x cos x  y( x )  sin x  y ( x )  (e  cos t ) y(t ) dt

 x  t
x cos x  y( x )  sin x  y ( x )   
or y ( x )  {e  cos t ) y(t )} dt
 0 x
d ( x) d (0) 
 (e x  cos x )  y( x )   (e0  cos 0) y(0) 
dx dx 
or y "( x )  x  cos x  y( x )  sin x  y '( x )  [0  (e x  cos x ) y( x )  1  0]
or y ( x )  sin x  y ( x )  e x y( x ) 
x (vi)
which is same as the given differential equation. To get the initial conditions,
we put x = 0 in Eq. (iv) and (v), and obtain
y(0) + 1 and y¢(0) = –1 (vii)
Thus Eqs. (vi) and (vii) together provide the original initial value problem.
EXERCISE 2.2
1. Convert the following initial value problems into integral equations.

d2 y
(a)  nxy  g( x ), y(0)
 1, y (0)
 0
dx 2
d2 y
(b) 2  y 0, y
(0) 0, y
(0) 0
dx
2. Convert y "( x )  3 y '( x )  2 y( x ) 4 sin x with initial conditions
y(0)  1, y (0)  2 into a Volterra integral equation. Also, find the
original initial value problem from the obtained integral equation.
3. Transform y   xy 1, y(0)  0, y(1)  1 into an integral equation.
4. Transform the boundary value problem y 
 y x, y
(0) 0, y
(1) 0 to
a Fredholm integral equation.

Answers:
x
1   ( x  t )[ g(t )  nt y(t )] dt
1. (a) y( x )  0
x
  ( x  t ) y(t ) dt
(b) y( x )  0
x
2. y(x) = 1 – x – 4 sin x + o {3 – 2(x – t)}y(t) dt
1 1 t 2 (1  x ), t  x
3. y( x
) x (1  x )   K ( x, t ) y(t ) dt; K ( x, t
) 
2 0
 xt (1  t ), t  x
x3 x 1 t , t  x
4. y( x )     K ( x, t ) y(t ) dt; K ( x, t )  
6 2 0
 x, t  x
2.4 BOUNDARY VALUE PROBLEM AND ITS CONVERSION TO

FREDHOLM INTEGRAL EQUATION
When an ordinary differential equation is given with the conditions involving
dependent variable and its derivatives at two different values of independent
variables, the problem under consideration is said to be a boundary value
problem.
The method of conversion of a boundary value problem to a Fredholm
integral equation can be made clear by the examples given hereunder.
EXAMPLE 2.6: Reduce the following boundary value problem into an

integral equation:
d 2u
0 with 
 lu  u(0) 0,
u(l) 0
dx 2
Solution:
The given conditions are
u(0) = 0 [i(a)]
u(l) = 0 [i(b)]
The given differential equation is
u(x) = –lu(x)
Integrating it with respect to x from 0 to x, we get
x x

0 u "( x) dx   l  u( x ) dx
0
x
or  l  u( x ) dx
u ( x )  u (0) 
0
x
Let u¢(0) = c so that u ( x ) c  l  u( x ) dx.
0

Again, on integration, we get
x x x x
0 u ( x) dx 
c  dx  l  u( x )(dx ) cx  l  u(t )(dt )2
2

0 0 0
x
or u( x )  u(0)  cx  l  ( x  t )  u(t ) dt
0
Now using Eq. [i(a)],

x
cx  l  ( x  t )u(t ) dt (ii)
u( x ) 
0
Now, to determine c, we use Eq. [i(b)] by taking x = l, so that
l
u(l) 
0cl  l  (l  t ) u(t ) dt
0
l l
l 0
 c (l  t )u(t ) dt
Now, Eq. (ii) can be expressed as

l l x
u( x )  x  (l  t ) u(t ) dt  l  ( x  t ) u(t ) dt
l 0 0
l l x (l  t ) x

u( x ) 0 l
u(t ) dt   l ( x  t ) u(t ) dt
0
x l x (l  t ) l l x (l  t ) x

or u( x ) 0 l
u(t ) dt  
x l
u(t ) dt   l( x  t ) u(t ) dt
0
x  x (l  t )  l x (l  t )
u( x ) l  
or   ( x  t ) u(t ) dt  l  u(t ) dt
0  l  x l
 x t (l  x ) l x (l  t ) 

or u( x ) l   u(t ) dt   u(t ) dt 
 0 l x l 
l
or u( x ) l  K ( x, t )  u(t ) dt (iii)
0
 t (l  x )
 l , if 0  t  x
or where, K ( x, t )   (iv)
 x (l  t ) , if x  t  l
 l
Equation (iii) is the required integral equation, whose kernel K(x, t) is defined
by Eq. (iv).
EXAMPLE 2.7: Transform the boundary value problem

d2 y

 y x, y
(0) 0, y 
(1) 0
dx 2
to a Fredholm integral equation. Also, recover the boundary value problem
from the integral equation.
Solution:
The given conditions are
y (0) = 0 [i(a)]
y¢(1) = 0 [i(b)]
Integrating the given differential equation with respect to x, we have
x x x

0 y ( x) dx  0 y( x) dx 
0 x dx
x2 x
or y ( x )  y 0)    y( x ) dx
2 0
Let y¢(0) = c, so that

 x2  x
y ( x )  c     y( x ) dx (ii)
 2 0
Integrating both sides with respect to x

x3 x
y( x )  y(0)  cx    y( x )(dx )2 (iii)
6 0
Using Eq. [i(a)] we express Eq. (iii) as

x3 x
y( x )  cx    y(t )(dt )2
6 0
 x3 x
y( x )  cx    ( x  t ) y(t ) dt [iv(a)]
6 0
Now, to determine c, we differentiate [iv(a)] and use Eq. [i(b)].

1 1
0 c    y(t ) dt
2 0

[Here, we apply Eq. (1.20)]
1 1
 c    y(t ) dt
2 0
and then, Eq. [iv(a)] is
3
 1 1  x x
y( x )  x     y(t ) dt     ( x  t ) y(t ) dt [iv(b)]
 2 0  6 0

y( x ) 
x x3

2 6
 0
x
x
1
 x
x.y(t ) dt   x. y(t ) dt   ( x  t ) y(t ) dt
0
3
x  3x x 1
y (
x)   t  y(t )dt   x  y(t ) dt
6 0 x
1 3 1
or y ( x ) ( x  3 x )   K ( x, t ) y(t ) dt (v)
6 0
 x, x  t
where, K ( x, t )   (vi)
t , x  t
Converse: We take the integral equation, i.e., Eq. [iv(b)].
x x3 1 x
y( x ) 
    x  y(t ) dt   ( x  t ) y(t ) dt (vii)
2 6 0 0
Differentiating both sides with respect to x, we have

1 x2 d 1 d x

y ( x )   
2 2 dx  0
x  y(t ) dt 
dx 0
( x  t ) y(t ) dt
1 x2 1 x
y ( x ) 
   y(t ) dt   y(t ) dt (viii)
2 2 0 0
Again, differentiating with respect to x, we get

d 1 d x

y ( x ) 
x 
dx 0
y(t )(dt ) 
dx 0
y(t )(dt )
x 
y ( x )  x  0   {y(t )} dt  y( x )
0 x
y ′′( x ) = x − 0 − y( x )

or y¢¢(x) + y = x (ix)
Also from Eqs. (vii) and (viii)
y(0) = 0 and y¢(1) = 0
So, Eq. (ix) is the original differential equation with the boundary conditions.
EXAMPLE 2.8: Transform y¢¢ + xy = 1, y(0) = 0, y(1) = 1 into an integral

equaton.

y¢¢ = 1 – xy (i)
with the boundary conditions
y(0) = 0 [ii(a)]
y(1) = 1 [ii(b)]
Integrating Eq. (i) with respect to x (x varying from 0 to x),
x x x

0 y ( x) dx  0 dx  0 x  y( x) dx
x
 x  x  y( x ) dx
y ( x )  y (0) 
0
Let y¢(0) = c, so that

x
y ( x )  c  x   x  y( x ) dx
0
Integrating once more with respect to x from 0 to x,

x x x
0 y ( x) dx  0 (c  x) dx  0 x  y( x)(dx)
2

x
 x2  x
or y( x )  y(0)   cx     t  y(t )dt )2
 2  0
0
1 x
or y( x ) 
 0 (2cx  x 2 )   ( x  t ) t y(t ) dt (iii)
2 0
Now, to determine c, we use Eq. [ii(b)], which is y(1) = 1.

1 1

1 (2c  1)   (1  t )  t  y(t ) dt
2 0

1 1
\   (1  t )  t  y(t ) dt
c (iv)
2 0
Putting the value of c in Eq. (iii), we get

2
1 1  x x
) x    (1  t )t  y(t ) dt  
y( x   t ( x  t )  y(t ) dt
2 0  2 0
x x 1 x
y( x ) (1  x )   xt (1  t )  y(t ) dt   xt (1  t )  y(t ) dt   t ( x  t )  y(t ) dt
2 0 x 0
x x 1
y( x ) (1  x )   t 2 (1  x )  y(t ) dt   xt (1  t )  y(t ) dt
2 0 x
x x
or y( x )  (1  x )   K ( x, t )  y(t ) dt (v)
2 0
t 2 (1  x ), when t  x
where, K ( x, t )   (vi)
 xt (1  t ), when t  x
Equation (v) is the required integral equation, whose kernel K(x, t) is defined
by Eq. (vi).
EXAMPLE 2.9: If m(x) is continuous and satisfies
1 (1  x ) x, 0  x  x
u
( x)  0 l  K ( x, x ) u(x ) dx , where
K ( x, x )  
(1  x )x , x  x  1
,
then prove that m(x) is also a solution of the boundary value problem
d 2u
 l  0, u(1)
u 0, u(0)  0
dx 2
Solution: Here, from the integral equation,
1
u( x ) l  K ( x, x )  u(x ) dx (i)
0
(1  x ) x, 0  x  x
where, K ( x, x )   (ii)
(1  x )x , x  x  1
We have to find the differential equation satisfying u(x) with the corresponding
boundary conditions. The kernel may be followed by Figure. 2.1 given below:
Figure 2.1 Example 2.9.
Equation (i) is now expressed as
u( x ) l   K ( x, x ) u(x ) dx   K ( x, x ) u(x ) dx 
x 1

 0 x 
x ) l   (1  x )x  u(x ) dx   (1  x ) x  u(x ) dx  (iii)

x 1
u(
 0 x 
Let x = 0, 1. We find u(0) = 0 [iv(a)]
and u(1) = 0 [iv(b)]
Now, differentiating Eq. (iii) with respect to x, we get [using Leibnitz’s
rule Eq. (1.20)]
du d x d 1

dx dx
0
l (1  x )x  u(x ) dx 
dx x
l x (1  x )  u(x ) dx
du x  dx d

dx
0 x l (1  x)x  u(x )dx  l (1  x)  x  u( x)  dx  0  dx (0)
1  d dx
 l x(1  x )  u(x ) dx  0  (1)  l x(1  x)  u( x) 
x x dx dx
du x 1
or 
dx 0 l (1)  x  u(x ) dx  x l (1  x )  u(x ) dx (v)
Differentiating Eq. (v) with respect to x using Leibnitz’s rule, we get
d 2u  x  dx 
 
 {lx u(x )} dx  l x  u( x )  0 
dx 2
 x
0 dx 
1  dx
 {l (1  x ) u(x )}dx  0  l (1  x )  u( x )
x x dx
d 2u d 2u
or 
 l x u( x )  l (1  x ) u( x )  0 (vi)
 lu 
dx 2 dx 2
which is the required equation with boundary conditions given by Eq. [iv(a)]
and [iv(b)].
EXERCISE 2.3
1. (a) If y¢¢(x) + ly(x) = 0, and y satisfies the end conditions y(0) = 0,

lx l x
y(l) = 0, show thaty( x )
l  0
(l  t ) y(t ) dt  l  ( x  t ) y(t ) dt .
0
(b) Show that the result of part (a) may be expressed as

l
y( x )  l  K ( x, t ) y(t ) dt
0

 t (l  x )
 l , when t  l
where, K ( x, t )  
 x (l  t ) , when t  x
 l
(c) Verify directly that the expression obtained in part (b) satisfies the
prescribed differential equations and end conditions.
2. Convert the problem y   l  y 0), y(p
y 0; y(0) ) y (p ) to an integral
equation.
Answer:
p
y( x ) l  K ( x, t )  y(t )dt
0
 (t  1)(p  x  1)
 ,0t x
p
where, K ( x, t )  
 ( x  1)(p  t  1) , x  t  p
 p
vvv
Chapter 3
Solution of Homogeneous
Fredholm Integral
Equations of the Second Kind
3.1 INTRODUCTION
In the previous chapters, we have learnt the basic terminology and suitability
of integral equations over differential equations, and then, the classification
of the integral equations. Although, there have been many developments in
the theory, the basic division as initial value problems into volterra integral
equations and boundary value problems into Fredholm integral equations is
a must to follow. As we approach to simplify these equations, it is found
convenient to focus the kernel, kind and the homogeneous nature of these
equations. In this chapter, we shall restrain ourselves only for the homogeneous
Fredholm integral equations of the second kind. The chapter begins with the
discussion of an essential part called eigenvalue, eigenfunction, and then
related theorems are explained.
3.2 CHARACTERISTIC VALUE* (OR EIGENVALUE) AND

CHARACTERISTIC FUNCTION (OR EIGENFUNCTION)
We consider the following homogeneous Fredholm integral equation of the
second kind:
b
u( x )  l  K ( x, t )u(t ) dt (3.1)
a
It is clear that u(x) = 0 will always satisfy Eq. (3.1), and we call such
u(x) = 0 as the trivial solution of Eq. (3.1). The values of parameter l for
which the integral equation [Eq. (3.1)] has non-trivial (non-zero) solutions
[u(x) ≠ 0] are known as eigenvalues of Eq. (3.1) or characteristic values of
kernel K(x, t). If for u(x) ≠ 0, there exists a continuous function f(x) in the
interval [a, b] such that
b
φ ( x )  l0  K ( x, t )φ (t ) dt (3.2)
a
*
It is also called characteristic number.
29
then f(x) is known as eigenfunction* of Eq. (3.1) corresponding to eigenvalue

l0.
Note:
1. If kernel K(x, t) is continuous for a  x  b, a  t  b, for the finite
values of a and b, then corresponding to every eigenvalue of l there
exists a finite number of linearly independent eigenfunctions; the number
of such functions is called index of the eigenvalue. Different eigenvalues
have different indices.
2. As seen above, along with f(x), cf(x) is also an eigenfunction for l0,
where c is an arbitrary constant (though not L.I.).
3. The number l = 0 is not taken as eigenvalue, since it provides trivial
solution.
4. A homogeneous Fredholm integral equation may not have eigenvalues
and eigenfunctions if the kernel is not symmetric.
EXAMPLE 3.1: Find the eigenvalue and eigenfunction of the homogeneous
1
integral equation g( x )  l 0 e g(t ) dt .
x t
Solution: The given integral equation is

1
g( x )  l e x  et g(t ) dt (i)
0
1 t
Let c e g(t )dt (ii)
0
so that Eq. (i) provides g(x) = lcex or g(t) = lcet (iii)

1 t
Then, Eq. (ii) providesc 0 e  l cet dt
l c  2t 1 l c 2
c 
e (e  1)
2  0 2
 l 
which is c 1  (e2  1)  
0
 2 
Now, for non-trivial solutions,
l 2
1 (e  1) 
0
2
2
 l 2
(iv)
e 1
* It is obvious that an eigenfunction will satisfy the integral equation.

Solution of Homogeneous Fredholm Integral Equations of the Second Kind 31
This is the eigenvalue of Eq. (i), and using Eq. (iii), the corresponding
eigenfunction is
2
 g( x ) 2
ce x e x
e 1
2c
by taking the constant 2 as unity.
e 1
EXAMPLE 3.2: Show that the homogeneous integral equation
1
f ( x ) l  (3 x  2) tf (t )dt

0
has no characteristic number and no eigenfunction.

1
( x ) l (3 x  2)  tf (t )dt (i)
f
0
1
Let c   tf (t )dt (ii)
0
so that f
( x ) l (3 x  2)c (iii)
or 
f (t ) l c(3t  2)
1
Then, c t  l c(3t  2)dt
0
1
  c l c t 3  t 2  
c 0
0
Thus, f(x) = 0 is the solution of Eq. (i) and we do not get any characteristic
number or eigenfunction.
We also see that kernel K(x, t) = (3x – 2)t is not symmetric, and in this
case, the kernel does not possess any characteristic number necessarily.
3.3 SOLUTION OF HOMOGENEOUS FREDHOLM INTEGRAL

EQUATION OF THE SECOND KIND WITH SEPARABLE (OR
DEGENERATE) KERNEL
Let the homogeneous Fredholm integral equation of the second kind be
b
u( x )  l  K ( x, t )u(t )dt (3.3)
a
where kernel K(x, t) is separable, which means it can be expressed as the

sum of the product of terms, each having functions of x and t separately. Let
n
K ( x, t )   fi ( x ) gi (t ) (3.4)
i 1
Then, Eq. (3.3) is
b n 
u( x )  l   fi ( x ) gi (t )  u(t )dt (3.5)
a
 i 1 
n b
u( x )  l  fi ( x )  gi (t )u(t )dt (3.6)
a
i 1
[By interchanging the order of integration and summation]
b
Let Ci g (t )u(t )dt , i
a i
1, 2, ..., n (3.7)
so that Eq. (3.6) reduces to

n
u( x )  l  Ci fi ( x ) (3.8)
i 1
We first find Ci as below:
Multiplying Eq. (3.8) by gi(x), (i = 1, 2, ..., n) successively and integrating
over the interval [a, b], we obtain
b n b
a g1 ( x )u( x ) dx  l  Ci  g1 ( x ) fi ( x ) dx [3.9(a)]
a
i 1
b n b
a g2 ( x)u( x) dx  l 
i 1
Ci  g2 ( x ) fi ( x ) dx [3.9(b)]
a
. . . .
. . . .
. . . .
b n b
a gn ( x)u( x) dx  l 
i 1
Ci  gn ( x ) fi ( x ) dx [3.9(n)]
a
b

We define a ji a g j ( x)  fi ( x)dx (i, j =1, 2, ..., n) (3.10)
Now using Eqs. (3.7) and (3.10), Eq. [3.9(a)] becomes
n
C1  l  Cia1i
i 1

or 
C1 l C1a11  C2a12    Cna1n 
or 1  la11  C1  la12C2    la1nCn 

0
Similarly, we derive other equations from Eq. (3.9), and finally, obtain
following simultaneous linear equations to establish Ci, which are
1  la11  C1  la12C2    la1nCn  0 


 la 21C1  1  la 22  C2    la 2 nCn 0

... ... ... ...
 (3.11) ...
... ... ...  ... ...

 la1nC1  la 2 nC2    1  la nn  Cn 
0

For this system of equations, let the determinant D(l) be
1  la11  la12    la1n

 la 21 1  la 22    la 2 n
D( l )  (3.12)
    
 la n1  la n 2   (1  la nn )
The eigenvalues of Eq. (3.3) are obtained by D(l) = 0, and we get the
maximum n eigenvalues.
3.4 ORTHOGONALITY OF TWO FUNCTIONS

Two functions f(x) and g(x) continuous for a  x  b, are said to be orthogonal if
b
a f ( x)g( x)dx  0 (3.13)
3.5 ORTHOGONALITY OF EIGENFUNCTIONS
Theorem: The eigenfunctions of a symmetric kernel corresponding to two
different eigenvalues are orthogonal.
Proof and explanation: Let the homogeneous Fredholm integral equation
of the second kind be
b
u( x )  l  K ( x, t ) u(t ) dt (3.14)
a
Here, kernal K(x, t) is symmetric and let f0(x) and f1(x) be eigenfunctions of
K(x, t) corresponding to eigenvalues l0 and l1(l0 ≠ l1). Then, we are required
to prove that f0(x) and f1(x) are orthogonal functions on the interval [a, b], i.e.,
b
a f0 ( x) f1 ( x)dx  0 (3.15)
Since f0(x) and f1(x) are eigenfunctions, by definition, these will satisfy
Eq. (3.14). Thus,
b
f0 ( x )  l0  K ( x, t )f0 (t ) dt (3.16)
a
and
b
f1 ( x )  l1  K ( x, t )f1 (t )dt (3.17)
a
Since kernel K(x, t) is a symmetric function, therefore

K ( x, t )  K (t , x ) (3.18)
Now, multiplying both sides of Eq. (3.16) by f1(x) and integrating with respect
to x over the interval [a, b], we get
b b
 b
a f0 ( x) f1 ( x)dx  l0 a f1 ( x) a K ( x, t ) f0 (t )dt  dx
b
 l0  f0 (t )
a 
a
b

K ( x, t ) f1 ( x )dx dt
(By changing the order of integration and now by Eq. (3.18)

b
 l0  f0 (t )
a  a
b

K (t , x ) f1 ( x )dx dt (3.19)
Now, interchanging the variables (t  x) in Eq. (3.17), we get

b
f1 (t )  l1  K (t , x )f1 ( x )dx
a
Using this value of f1(t) in Eq. (3.19), we express

b b  f1 (t ) 
a f0 ( x) f1 ( x) dx  l0 a f0 (t )  l1 
 dt
or
b b
l1  f0 ( x ) f1 ( x ) dx  l0  f0 ( x ) f1 ( x )dx
a a
By the property of definite integrals,

b
 (l1  l0 )  f0 ( x ) f1 ( x ) dx 
0
a
Now, since l1 ≠ l0, (l1 – l0) ≠ 0, and hence,

b

a f0 ( x) f1 ( x) dx  0
This proves the proposition.
3.6 REAL EIGENVALUES

Theorem: The eigenvalues of a symmetric kernel are real.
Proof: We take the homogeneous Fredholm integral equation of the second
kind.
b
u( x )  l  K ( x, t ) u(t )dt (3.20)
a
Let l0 = a + ib be one eigenvalue and f0(x) = u + iv be the corresponding
eigenfunction. Then,
l 0 a  ib ,  l0  l 0  2ib , 
and  2 2
(3.21)
f 0 ( x ) u  in  f0 ( x )  f 0 ( x )  u  v 
where, a bar denotes the complex conjugate. By the definition of eigenfunction,
f0(x) and f0 ( x ) will satisfy Eq. (3.20), and we get
b
f0 ( x )  l0  K ( x, t )f0 (t )dt [3.22(a)]
a
and
b
f 0 ( x )  l 0  K ( x, t ) f 0 (t )dt [3.22(b)]
a
Now, multiplying both sides of Eq. [3.22(a)] by f0 ( x ) and integrating with

respect to x over the interval [a, b], we find
b b

a f0 ( x) f 0 ( x)dx  l0 a f 0 ( x) a K ( x, t ) f0 (t )dt
b

dx

b
a f0 ( x) f 0 ( x)dx  l0 a f0 (t )
b
K ( x, t ) f 0 ( x )dx dt a
b

By changing the order of integration, and now since K(x, t) = K(t, x), we have

b b

a f0 ( x) f 0 ( x)dx  l0 a f0 (t ) a K (t, x) f 0 ( x)dx
b
 dt (3.23)
Now, interchanging the variables (x  t), Eq. [3.22(b)] takes the shape as
b
f 0 (t )  l 0  K (t , x ) f 0 ( x )dx
a
and then, Eq. (3.23) is
b b 1 
a f0 ( x)f 0 ( x)dx  l0 a f0 (t )  l 0 f 0 (t ) dt

b b
or l 0  f0 ( x ) f 0 ( x )dx  l0  f0 ( x ) f 0 ( x )dx
a a
(Using the property of definite integrals)
l 
b
or 0  l0 f0 ( x ) f 0 ( x )dx 
0 (3.24)
a
Using Eq. (3.21), we get
 
b
2ib  u2  v 2 dx 
0
a
Since f0(x) is an eigenfunction which is not zero, we infer
a u 
2 b
 v 2 dx  0

and thus, we conclude b = 0, which proves the proposition.
EXAMPLE 3.3: Show that the homogeneous integral equation
 
1
f ( x ) l  t x  x t f (t )dt

0
has no real eigenvalue and no eigenfunction.
Solution: The given equation is
 
1
f ( x ) l  t x  x t f (t )dt

0
1 1
or f ( x ) l x  tf (t )dt  l x  t f (t )dt (i)
0 0
1
Let c1   tf (t )dt (ii)
0
1
and c2   t f (t )dt (iii)
0
so that Eq. (i) is 

f ( x ) l c1 x  l c2 x
Thus, 
f (t ) l c1 t  l c2 t (iv)
0 t l c1 
1

and then, by Eq. (ii), c1 t  l c2 t dt
1 1
 t 5/2   t3 
c1 l c1    l c2  
 5/2 0  3 0
 2l  l
 c1 1   c 0 (v)
 5  3 2
Similarly, by Eq. (iv) and (iii)
 
1

c2 0 t l c1 t  l c2 t dt
1 1
 t2   t 5/2 
c2 l c1    l c2  
 2 0  5 / 2 0
l  2l 
  c1  1   c2 
0 (vi)
2  5 
The system of Eqs. (v) and (vi) will have a non-zero solution if
2l l
1
5 3
 D( l )  0
l 2l
 1
2 5
4l 2 l 2 l2
 1  0 1  0  l   i 150
25 6 150
which means that the given integral equation does not possess any real
eigenvalue.
EXAMPLE 3.4: Solve the following homogeneous Fredholm integral

equation of the second kind:
2p
g(s) l  sin(s  t ) g(t )dt
0
2p

Solution: g(s) l 
0
sin s cos t  cos s sin t  g(t )dt
2p 2p

g(s) l sin s  cos t g(t )dt  l cos s  sin t g(t )dt (i)
0 0
2p
Let c1   cos t g(t )dt (ii)
0
2p
and c2   sin t g(t )dt (iii)
0
so that Eq. (i) reduces to
g(
s) l sin s  c1  l cos s  c2 
or  (iv)
 g(t ) l c1 sin t  l c2 cos t 
Now, by Eq. (ii),
2p
c1  cos t  l c1 sin t  l c2 cos t  dt
0
2p 2p
l c  cos 2t  lc  sin 2t 
or c1  1     2 t  2 
2  2 0 2  0
l c2
or c1 0  [2p ]  c1  lp c2 0 (v)
2
Similarly, by Eqs. (iv) and (iii),
2p

c2 0 sin t{l c1 sin t  l c2 cos t} dt
2p 2p
l c1  sin 2t  l c  cos 2t 
c2   t   2 
2  2 0 2  2 0

or c2 l c1p  l c1p  c2 0 (vi)
For non-zero solution of this system of linear equations,
1  lp
D( l )  0   0  1  l 2p 2  0
lp 1

1
or l 
p
1 1
Hence, the eigenvalues are given by l1  and l2   .
p p
Now, determining eigenfunction,
1
(a) For l  , Eqs. (v) and (vi) are c1 – c2 = 0; hence, from Eq. (iv),
p
1 1
g(s) c1 sin s  cos s  c1
p p
c1

g(s) (sin s  cos s)
p
c
or g(s) (sin s  cos s) , taking 1  1
1 p
(b) For l   , Eqs. (v) and (vi) are c1 + c2 = 0,
p
Hence, from Eq. (iv)
1 1
g(s) c1 sin s  ( c1 ) cos s
p p
c1

g(s) (sin s  cos s)
p
c1
or 
g(s) (sin s  cos s) , taking 1
p
1
Finally, for the eigenvalue l  , the eigenfunction is g(s) = sin s + cos s
1 p
and for eigenvalue l   , the eigenfunction is g(s) = sin s – cos s.
p
EXAMPLE 3.5: Find the eigenvalues and eigenfunctions of the following
homogeneous integral equation:
p
( x ) l  (cos2 x  cos 2t  cos3 x  cos3 t ) f (t ) dt
f
0
Solution: The given integral equation can be expressed as

x p
f ( x ) l cos2 x  cos 2t  f (t )dt  l cos3 x  cos3 t  f (t ) dt (i)

0 0
x
Let c1   cos 2t f (t )dt (ii)
0
x
and c2   cos3 t φ (t ) dt (iii)
0
so that Eq. (i) is
f ( x ) l c1 cos2 x  l c2 cos3 x

or f (t ) l c1 cos2 t  l c2 cos3t (iv)
Then, by Eq. (ii),
p
c1  cos 2t  (l c1 cos2 t  l c2 cos3t ) dt
0
or  x  x
c1 1  l  cos 2t  cos t dt  l c2  cos 2t  cos3t dt 
2
0 .
 0  0
 p  1  cos 2t   p1
or c1 1  l  cos 2t    dt   l c2  (cos 5t  cos t ) dt 
0
 0  2   0 2
p
 p  cos 2t 1    sin 5t sin t 
or c1 1  l    (1  cos 4t )  dt   l c2    
0
0  2 4    10 2 0

  sin 2t t sin 4t 
p
or c1 1  l      l c2  0 0
  4 4 16 0
 lp 
or c1 1   0  c2 0 (v)
 4 
Similarly, by Eqs. (iii) and (iv),
p
c2  cos3 t  (l c1 cos2 t  l c2 cos3t )dt
0
p
0 cos
5
We find that t0
p p 1
0 cos 0
3
and t  cos3t
dt (cos3t  3cos t )  cos3t dt
4
1 p 1  cos 6t 3 p


4 0 2
dt   cos t cos3t dt
4 0
p
1  sin 6t  3 p
  t     (cos 4t  cos 2t )  dt
8 6 0 8  0 

p
p 3  sin 4t sin 2t  p
    
8 8 4 2 0 8

p  lp 
\ c2 0  c1  l
c2 or 0c1  c2 1   0 (vi)
8  8 
For non-zero solution of system of Eqs. (v) and (vi), D(l) = 0.
lp
1 0
4
0
lp
0 1
8
or  lp   lp 
 1    1 0
 4  8 
8 4
 l ,
p p
which are the eigenvalues.
Now, determining eigenfunctions,
(a) For l  4/p ; by Eq. (v), c1  0  c2  0  0 

 4 p    c2 
0, c1 is arbitrary
and by Eq. (vi), c2 1    0 
 p 8  
Hence, by Eq. (iv),
f(x) = lc1 cos2 x + 0
4
or f ( x)  c1 cos2 x
p
 4c 
or f(x) = cos2 x  taking 1  1
 p 
 p 8 

(b) For l 8/p ; by Eq. (v), c1 
1  4  p   0  c1 0 
  
  c1 
0 and c2 is arbitrary
 p 8  
and by Eq. (vi), 0c1  c2 1    0 
 8 p  
Then [by Eq. (iv)] the corresponding eigenfunction is
8 8
f ( x)   0  cos2 x   c2 cos3 x
p p
 8c2 
f(x) = cos 3x  taking  1
p 
EXERCISE 3.1
Find the eigenvalues and the corresponding eigenfunctions of the following

integral equations:
1
x ) l  sin p x  cos p t  u(t ) dt
1. u(
0
1

2. f ( x ) l  (2 xt  4 x 2 ) f (t ) dt
0
2 1
g(s) l   st   g(t ) dt
3.  1  st 
1
( x ) l  (5 xt 3  4 x 2 t  3 xt ) f (t ) dt
4. f
1
p
5. g( x ) l 0 [cos x cos 2t  cos3 x cos t ] g(t ) dt
2 3

Answers:
1. No eigenvalue
2.
3,  3, ( x  2 x 2 )
1
 
 1  1
3. 17  265 , s  2  2732   , s  0.4399  
2  s  s
1 3x
4. ) x2 
, f ( x
4 2
4 8 2

5. 
, , g1 ( x ) cos x, g2 ( x ) cos3 x
p p
Chapter 4
Fredholm Integral Equations

with Separable Kernels
4.1 INTRODUCTION
Like the previous chapter, this chapter is also devoted to Fredholm integral
equations of the second kind. But, now the equation is not homogeneous,
i.e., F(x) ≠ 0 in general. Furthermore, the nature of kernel is separable. This
chapter shows that the role of characteristic function with determinant D(l) is
of vital importance. All interrelated possible cases are included with suitable
examples.
4.2 SOLUTION OF FREDHOLM INTEGRAL EQUATION OF

THE SECOND KIND WITH SEPARABLE (OR DEGENERATE)
KERNEL
We consider the following integral equation:
b
( x ) F ( x )  l  K ( x, t )u(t )dt (4.1)
u
a
where, kernel K(x, t) is separable, and therefore, we express

n
K ( x, t )   fi ( x ).gi (t ) (4.2)
i 1
which on substituting in Eq. (4.1) provides

b n 
( x ) F ( x )  l   fi ( x ).gi (t )  u(t )dt
u
a
 i 1 
n b
or ( x ) F ( x )  l  fi ( x )  gi (t )u(t )dt (4.3)
u
a
i 1
(upon interchanging the order of integration and summation)
41
b
We prescribe  gi (t
)u(t )dt c
i, (i 1, 2, ..., n) (4.4)
a

which shapes Eq. (4.3) as below:
n
( x ) F ( x )  l  ci fi ( x ) (4.5)
u
i 1
Now, for determining ci’s, we multiply Eq. (4.5) by g1(x), g2(x), ..., gn(x)
successively and integrate over the interval [a, b], and frame
b b b
∫a g1 ( x )u( x )dx = ∫a F ( x ) g1 ( x )dx + λ ∑ ci ∫a g1 ( x ) ⋅ fi ( x )dx [4.6(a)]
b b b
a g2 ( x)u( x)dx  a F ( x)g2 ( x)dx  l  ci a g2 ( x)  fi ( x)dx [4.6(b)]
………………………………………………………………
………………………………………………………………
………………………………………………………………
b b b
a gn ( x)u( x)dx  a F ( x)gn ( x)dx  l  ci a gn ( x)  fi ( x)dx [4.6(n)]
We now define
b
a ji   g j ( x ) fi ( x )dx , (i, j = 1, 2, ..., n) (4.7)
a
b
and b j   g j ( x )F ( x )dx , (j = 1, 2, ..., n) (4.8)
a
Now, using Eqs. (4.4), (4.7), (4.8) for [4.6(a)], we obtain
n
c
1 b1  l  cia1i
i 1
or c1  b1  l[c1a11  c2a12   cna1n ]

or 1  la11  c1  la12 c2   la1n cn 
b1 [4.9(a)]
A similar simplification for Eqs. [4.6(b), ..., 4.6(n) will provide
 la 21c1  1  la 22  c2   la 2 n cn  b2 [4.9(b)]
..... ..... ..... ..... .....
 la n1c1  la n 2 c2   1  la nn  cn  b n [4.9(n)]
The determinant D(l) of this system of linear equations, i.e., from
Eq. [4.9(a)] to [4.9(n)] gives
1  la11   la12  la1n
D( l ) 
 la 21 1  la 22   la 2 n

 la n1 1  la nn 
 la n 2
D(l) = 0 is a polynomial of utmost degree n in l. D(l) does not vanish
identically, since D(0) = 1. We consider the following three cases:
Fredholm Integral Equations with Separable Kernels 43
Case 1: When F(x) = 0, then by Eq. (4.8), bj = 0, which makes the system
[Eq. (4.9)] a homogeneous linear equation. We consider the following two
situations:
(a) If D(l) ≠ 0, then we get a unique zero solution, i.e., c1 = c2 = ...
= cn = 0 for Eq. (4.9). So Eq. (4.1) has the solution u(x) = 0.
(b) If D(l) = 0, those values of l for which D(l) = 0 are the eigen values
and any non-zero solution of homogeneous [ F(x) = 0] Fredholm
integral equation is known as eigenfunction of the integral equation.
Case 2: When F(x) ≠ 0, but
b
a g j ( x)F ( x)  0, j = 1, 2, ..., n,
i.e., F(x) is orthogonal [see Eq. (3.3)] to all functions gj(x), then by Eq. (4.8),
bj = 0. (j = 1, 2, ..., n), and hence, the system [Eq. (4.9)] reduces to
homogoneous linear equations. We find the following two situations:
(a) If D(l) ≠ 0, then the unique zero solution is ci = 0 for Eq. (4.9),
which provides the solution of Eq. (4.1) as u(x) = F(x), [here, we
use Eq. (4.5)).
(b) If D(l) = 0, then the system [Eq. (4.9)] provides infinite non-zero
solutions, and thus Eq. (4.1) has infinite non-zero solutions. The
resulting solution is the sum of F(x) together with the arbitrary
multiples of eigenfunctions.
Case 3: When at least one of bj is not zero, we have the following two
situations:
(a) If D(l) ≠ 0, we get a unique non-zero solution of Eq, (4.9), and
hence, a unique non-zero solution of Eq. (4.1) is obtained by Eq. (4.5).
(b) If D(l) = 0, we get either no solution or infinite solutions, and thus,
Eq. (4.1) has no solution or infinite solutions.
1
(s) f (s)  l  st g(t )dt .
EXAMPLE 4.1: Solve g
0

1
(s) f (s)  l s  t g(t )dt (i)
g
0
1
Let c   t g(t )dt (ii)
0
so that g
(s) f (s)  l sc (iii)
or g
(t ) f (t )  l tc (iv)
1

by which Eq. (ii) provides c 0 t[ f (t )  ltc]dt
1
1  t3 
or c  t f (t )dt  l c  
0
 3 0
1 lc
or  c  t f (t )dt 
0 3
 l 1 3 1
or c 

1 
3 0 t f (t )
dt  c
3  l 0
t f (t )dt , l  3
Then, by Eq. (iii), the solution Eq. (i) is

3l s 1
3  l 0
g
(s) f (s)  t f (t )dt

p
( x ) cos x  l  sin x f (t )dt.
EXAMPLE 4.2: Solve f
0

p
( x ) cos x  l sin x  f (t )dt (i)
f
0
p
Let c   f (t )dt (ii)
0
so that f
( x ) cos x  l sin x.c (iii)
Hence, f
(t ) cos t  l sin t .c (iv)
p

c 0 (cos t  l c sin t )dt
or 
c sin t p0  l c   cos t p0
 0  l c[ ( 1)  1]
c 2l c
c(1  2l ) 
0
1
or 
 c 0 if l 
2
1
By Eq. (iii), the solution is f(x) = cos x, provided l  .
2
1
1   (1  e x t ).y(t )dt.
EXAMPLE 4.3: Solve y( x ) 
0

1 1
1   y(t )dt  e x  et y(t )dt (i)
y( x ) 
0 0
1
Let c1   y(t )dt (ii)
0
1
and c2   et y(t )dt (iii)
0
so that Eq. (i) is y( x ) 1  c1  e x c2 (iv)

Then, y(t ) 1  c1  et c2 (v)
1
0 [1  c1  c2 e ] dt
t
Therefore, c1 
1
or c1 t  c1t  c2 et 
0
or c1 1  c1  c2 (e  1)
1
or c2  (vi)
Now, using Eq. (v), 1e
1 t
0 e [1  c1  c2 e ] dt
t
c
2

1
 c 
c2 =  et  c1et  2 e2t 
 2 0
c2 2
c2 (e  1)(1  c1 )  (e  1)
2
Putting for c2,
 1   (e  1) 
2
  e   1  (e  1)(1  c1 )
1  2 

(e2  2e  3)
or c1   (vii)
2(e  1)2
Putting c1 and c2 is Eq. (iv), we get the required solution.
EXAMPLE 4.4: Show that the integral equation

1 2p
p 0
g(s
) f (s)  sin(s  t )  g(t )dt

possesses no solution for f(s) = s, but it possesses infinitely many solutions
when f(s) = 1.
1 2p
f (s)   (sin s cos t  cos s sin t ).g(t )dt
g(s) 
p 0
1 2p cos s 2p
sin s   cos t g(t )dt 
p 0
or g(s) f (s)  sin t  g(t )dt (i)
p 0
2p
Let c1   cos t g(t )dt (ii)
0
2p
and  c2 0 sin t  g(t ) dt (iii)
so that Eq. (i) reduces as
sin s cos s
g(s) f (s)  c1  c2 (iv)
p p
As per the given value of f(s), we take the following two cases:
Case 1: When f(s) = s, so that
c c
g(s)  s  1 sin s  2 cos s (v)
p p
c1 c2
Hence, g(t ) 
t  sin t  cos t (vi)
p p
which provides by Eq. (ii)
2p  c c 
c1 0 (cos t ) t  1 sin t  2 cos t  dt
 p p 

2p c1 2p c 2p
0 t cos t dt 
or c1 
p 0
sin t cos t dt  2
p 0 cos2 t dt
2p 2p
2p c1 1  1  c  sin 2t 
c1 = t sin t 0  
2p
sin t dt  .   cos 2t   2 t  2 
0 p 2 2 0 2p  0
c2
or c1  0    cos t 0  0 
2p
[2p  0]
2p
or c1 – c2 = 0 (vii)
Similarly, by Eqs. (iii) and (iv),
2p  c c 
 c2  sin t  t  1 sin t  2 cos t  dt

0  p p 
2p c 2p c 2p
or c
2 0 t sin t dt  2p1 0 (1  cos 2t )dt  2p2 0 sin 2t dt
2p 2p
2p c1  sin 2t  c2  cos 2t 
 t cos t 0  
c2 
2p
(  cos t )dt 
2p t  2   2p  2 

0  0  0
c1
 2p  sin t 0  2p  0  0 (viii)
2p
or c1 
2p
or c1  c2 
2p
Looking at Eqs. (vii) and (viii), we find that this system is inconsistent,
and hence, possesses no solution.
Case 2: When f(s) = 1, then by Eqs. (v) and (vi),
c c
1  1 sin s  2 cos s (ix)
g( s ) 
p p
c c
Hence, 1  1 sin t  2 cos t (x)
g(t ) 
p p
2p  c c 
Then by Eq. (ii), 
c1 0 cos.t 1  1 sin t  2 cos t  dt (xi)
 p p 
which provides c1 = c2.
Again, by Eqs. (iii) and (x) ,
2p  c c 
c
2 0 sin t 1  1 sin t  2 cos t  dt (xii)
 p p 
which provides c1 = c2.

Thus, Eqs. (xi) and (xii) provide c1 = c2 = c0 (arbitrary constant), and
hence, by Eq. (ix), we have
c
g(s)  1  0 (sin s  cos s)
p
or g(s)  1  c(sin s  cos s)
It is the required solution. Clearly, c can have arbitrary value; hence, in case
when f(s) = 1, we find infinitely many solutions to Eq. (i).
EXAMPLE 4.5: Solve the integral equation
 x cos t  t 
p
f ( x)  l  2
sin x  cos x sin t f (t )dt 
x
p

p p p
x  l x  cos t  f (t )dt  l sin x  t 2f (t )dt  l cos x  (sin t )  f (t )dt (i)
f ( x) 
p p p
π
Let c1 = ∫ φ(t ) ⋅ cos t dt (ii)
−π
p
p f (t )  t
2
c2 dt (iii)
p
c3 p f (t )  sin t dt (iv)
so that Eq. (i) can be expressed as
f ( x) 
x  l c1 x  l c2 sin x  l c3 cos x (v)
\ f (t ) 
t  l c1t  l c2 sin t  l c3 cos t (vi)
p
Thus c1  p (t  l c1t  l c2 sin t  l c3 cos t ) cos t dt
p p p
(1  l c1 )  t cos t dt  l c2  sin t cos t dt  l c3  cos2 t dt
or c1  p p p
p 1
c1  0  0  2 l c3  (1  cos 2t )dt
0 2
a
[ f ( x )dx  0 if f ( x ) is an odd function]
a
p
 sin 2t 
or c l c3 t   c1  lp c 0 (vii)
2 0
1 3

By Eqs. (iii) and (vi),
p
p [t  l c1t  l c2 sin t  l c3 cos t ]t
2
c2  dt

p p p
 (1  l c1 )t dt  l c2  t sin tdt l c3  t cos t dt
3 2 2
or c2 
p p p
 
c2  0  0  2l c3  t 2 sin t   2t sin t dt 
p p

 0 0 
4l c3 t (  cos t )0   (  cos t )dt 

p p
c2 
 0 
c2 4l c3 [p  0]  c2  4pl c3  (viii)
0
Similarly, by Eqs. (iv) and (vi),
p

c3  p [t  l c1 t  l c2 sin t  l c3 cos t ]  sin t dt
p p
2(1  l c1 )  t sin tdt  2l c2  sin 2 tdt  0
c3 
0 0
 2l c2 
c3  2(1  l c1 ) t (  cos t )0   (  cos t )dt   
p p p
 0 
  2 0 (1  cos 2t )dt 
p
 1 
c3  2(1  l c1 ) p  0   l c2 t  sin 2t 
 2 0
or c3  2p (1  l c1 )  l c2p
or 2pl c1  l c2p  c3 2p (ix)
Solving Eqs. (vii), (viii) and (ix), we get
2p 2 l 8p 2 l 2p
c1  , c2  , c3 
1  2l p 2 2 1  2l p 2 2
1  2 l 2p 2
Putting these values of c1, c2 and c3 in Eq. (v), we get the required solution as
2pl
f ( x) 
x (lp x  4 lp sin x  cos x )
1  2 l 2p 2

2p
f ( x) l 
g( x )  sin( x  t ) g(t ) dt
0
and discuss all possible cases.
2p
f ( x) l 
g( x )  sin( x  t ) g(t ) dt (i)
0
2p
g(x)  f ( x )  l  [sin x cos t  cos x  sin t ] g(t ) dt
0
2p 2p
g(x) f ( x )  l sin x   cos t  g(t ) dt  l cos x  sin t  g(t ) dt
0 0
or g( x ) 
f ( x )  (l sin x )c1  (l cos x )c2 (ii)
2p
where, c1   cos t g(t ) dt (iii)
0
2p
and c2 0 sin t  g(t ) dt (iv)
Now, from Eq. (ii),
g(t ) 
f (t ) (l sin t )c1  (l cos t )c2 (v)
Putting Eq. (v) in Eq. (iii), we have
2p

c
1 0 cos t[ f (t )  (l sin t )c1  (l cos t )c2 ] dt
2p 2p sin 2t 2p 1
c1  0 cos t  f (t ) dt  l c1 
0 2
dt  l c2 
0 2
(1  cos 2t ) dt
2p
c1 0 cos t  f (t ) dt  0  l c2  p
2p
c1  lp c2 0 cos t  f (t ) dt (vi)
Similarly, by Eqs. (v) and (iv),

2p
 lp c1 
c2 0 sin t  f (t ) dt (vii)
From Eq. (vi) and (vii),

1  lp
the corresponding D( l 
)  (1  l 2p 2 ) (viii)
 lp 
1
\ D( l ) 
(ix)0l 
p
Case 1: When f(x) = 0, then Eq. (i) reduces to
2p
g( x ) l 
 sin( x  t )  g(t )dt
0
(a) If D(l) ≠ 0, then we shall have a unique zero solution, i.e., c1 = 0

= c2. So, Eq. (i) has the solution g(x) = 0.
1
(b) If D(l) = 0 (case of non-trivial solution), i.e., if l   , then the
1 p
eigenvalues are l   .
p
1 1
We now find eigenfunctions corresponding to l  , .
p p
1
For l  , using Eq. (vi) and (vii) with f(x) = 0, these equations give
p
c1 – c2 = 0, and therefore, Eq. (ii) becomes
c
g( x )  1 (sin x  cos x )  A(sin x  cos x )
c p
where, A  1 is an arbitrary constant.
p
Thus, (sin x + cos x) or any non-zero constant multiple of (sin x + cos x)
1
will be the eigenfunction corresponding to eigenvalue l  .
1 p
Similarly, corresponding to l  , Eq. (vi) and (vii) with f(x) = 0 give
p
c1 + c2 = 0, and therefore, Eq. (ii) becomes
c
g( x )  2 (sin x  cos x )  B(sin x  cos x )
p
c2
where, B = is an arbitrary constant.
π
Thus, (sin x – cos x) or any non-zero constant multiple of (sin x – cos x)
1
will be the eigenfunction corresponding to eigenvalue l   .
p
Case 2: When f(x) ≠ 0, but
2p 2p
0 cos t  f (t ) dt 
0 and 0 sin t  f (t ) 
0
i.e., f(t) is orthogonal to cos t and sin t, then Eqs. (vi) and (vii) provide
homogeneous linear equations. We have the following two situations:
(a) If D(l) ≠ 0, Eqs. (vi) and (vii) provide c1 = 0 = c2. So, the solution
of Eq. (i) is g(x) = f(x).
1
(b) If D(l) = 0, then l   and as found in Case 1, we have infinite
p
non-zero solutions. The resulting solution of (1) is g(x) = f(x) + the
1
arbitrary multiples of (sin x + cos x) if l  , and (sin x – cos x) if
1 p
l  .
p
2p 2p
Case 3: When at least one of the  cos t  f (t )dt  0 or 0 sin t  f (t )dt  0
0
we have the following two situations:
(a) If D(l) ≠ 0, we will have a unique non-zero solution of Eq. (i).
1
(b) If D(l) = 0 and l  , then Eq. (vi) and (vii) become
p
cos t  f (t ) dt 
2p

c1 
c2
 [x(a)] 0
2p

  sin t  f (t )dt 
c1  c2  [x(b)]
0 
When l = –1/p, then Eqs. (vi) and (vii) provide
cos t  f (t ) dt 
2p


c1 
c2 0
2p
 [xi(a)]
c2  sin t  f (t )dt 
c1  [xi(b)]
0 
Now, Eqs. [x(a)] and [x(b)] are incompatible unless the function f(t) satisfies
the condition.
2p 2p

0   sin t  f (t ) dt
cos t  f (t ) dt 
0
2p
or 0 (cos t  sin t )  f (t ) dt 
0 (xii)
Similarly, Eqs. [xi(a)] and [xi(b)] are incompatible unless the function f(t)
satisfies the condition
2p
0 (sin t  cos t )  f (t ) dt 
0 (xiii)
When Eqs. (xii) and (xiii) are satisfied, the Eqs. (x) and (xi) become
redundant and we have infinitely many solutions.
1
Another case: When l  and Eq. (xii) is satisfied, then by Eq. (x)
p
2p
c2  
c1  cos t  f (t ) dt
0
Now, putting it in Eq. (ii).

c2 sin x sin x 2p c

g( x ) f ( x ) 
p

p 0 cos t  f (t ) dt  2 cos x
p
sin x 2p
p 0
g(x) f ( x )  cos t  f (t ) dt  A(cos x  sin x ) (xiv)
where, A = c2/p is an arbitrary constant.

2p
Thus, if l = 1/p and  (cos x  sin x ) f ( x ) dx  0 , Eq (i) possesses
0
infinitely many solutions.
Similarly, when l = –1/p, and Eq. (xiii) is satisfied, then by Eq. (xi),
2p
c1 c2   cos t  f (t ) dt
0
Putting it in Eq. (ii), we get
sin x 2p
x 0
g( x ) f ( x )  cos t  f (t ) dt  B(sin x  cos x ) (xv)
2p
Thus, if l = –1/p and 0 (sin x  cos x ) f ( x ) dx 
0 , Eq. (i) possesses
infinitely many solutions.
EXAMPLE 4.7: Solve the following integral equation and discuss its all
possible cases:
1
φ ( x ) F ( x )  l  (1  3 xt )φ (t )dt
0
Solution:
1
The given integral equation is f(x) = F(x) + l  (1  3 xt ) f (t )dt (i)
0
It may be expressed as
1 1
F ( x )  l  f (t ) dt  3 x l  t φ (t )dt (ii)
φ( x) 
0 0
1
Let c1   φ (t )dt (iii)
0
1
and c2   t φ (t )dt (iv)
0
so that Eq. (ii) shapes as below:
φ ( x )  F ( x )  l c1  3 x l c2 (v)
so that φ (t )  F (t )  c1l  3l c2 t (vi)
1
Thus, c
1 0 [F (t )  c1l  3l c2t ]dt
1 1
or 
c1 0 F (t )dt  c1l  3l c2  2 (vii)
Similarly, by Eqs. (iv) and (vi),
1


c2 0 t[F (t )  c1l  3l c2t ]dt
1 1 1
or c2 0 tF (t )dt  c1l  2  3c2 l  3 (viii)
Equations (vii) and (viii) may be expressed as
3l 1 1
(1  l )c1  
c
2 2 0 F (t )
dt  F ( x )dx (ix)
0
l 1 1
and 
2
c1  (1 
l )c2 0 t F
(t )dt  x F ( x )dx (x)
0
The determinant
1  l 3l /2 1
(l )
D  (4  l 2 ) (xi)
l / 2 1  l 4
Case 1: A unique solution of the system of equations i.e., Eqs. (ix) and
(x) will exist if and only if D(l )  0  l  2 . As discussed in Section 4.2
(Case 3), the values of c1 and c2 can be determined by solving this system,
which in turn, after putting in Eq. (v) gives the solution of integral equation,
i.e., Eq. (ii) Particularly, if F(x) = 0 and l ≠ ±2, we derive c1 = 0 = c2,
which leads to trivial solution f(x) = 0. Clearly, the numbers l = ±2 are the
eigenvalues for the problem.
If l = 2, Eqs. (ix) and (x) shape as
1 
c1  3c2 0 F ( x)dx  1
and
1
  0 (1  x )F ( x )dx 
0
c1  3c2 0 xF ( x ) dx 

This means that for l = 2, the system of equations becomes incompatible
1
(possessing no solution) unless  (1  x )F ( x ) 
0 . If this is true, then we get
0
same set from both the equations, which means one relation in c1 and c2, and
hence, infinitely many solutions will exist.
If l = –2, a similar argument follows.
Case 2: Let F(x) = 0, then Eq. (i) is homogeneous integral equation;
1
f ( x ) l  (1  3 xt ) φ (t ) dt (xii)

0
If l ≠ ±2, then Eq. (i) has the trivial solution f(x) = 0, as mentioned in
Case 1.
For non-trivial solution l = ±2 are the eigenvalues, we find eigenfunction
for l = 2 first. Eqs. (ix) and (x) for F(x) = 0 reduce to
c1 = 3c2
and then, by Eq. (v),
f(x) = 2(3c2 – 3xc2) = A(1 – x)
where, A = 6c2 is an arbitrary constant.
Thus, in this case, for l = 2, the eigenfunction is A(1 – x).
Similarly, for l = –2, Eq. (ix) and (x) provide c1 = c2.
Again, by Eq. (v),
f ( x) 
2c1 (1  3 x ) 
B(1  3 x )
where B = –2c1 is an arbitrary constant.
Clearly, in this case, when l = –2, the eigenfunction is B(1 – 3x).
Case 3: When F(x) ≠ 0, then the given integral equation, i.e., Eq. (i) is
non-homogeneous. We consider the following three situations.
(a) When l ≠ ±2: This situation has already been dealt in Case 1.
(b) When l = 2: The compatibility is there if F(x) is orthogonal to
1
(1 – x). We refer back to Case 1 [if l = 2, and  (1  x )F ( x ) dx 
0]
0
and we have found
1
c1 3c2   F ( x ) dx

0
which by Eq. (v) gives

1
f ( x ) F ( x )  l[3c2   F ( x ) dx ]  3 xc2 l
0
1
or f ( x )  F ( x )  2  F ( x ) dx  A(1  x ) (xiii)
0
where, A = 6c2
1
Finally, if l = 2 and  (1  x ) F ( x )  0 , Eq. (xiii) gives infinitely
0
many solutions.
(c) When l = –2: As before, the compatibility is there if F(x) is orthogonal
with (1 – 3x).
1
i.e., 0 (1  3x) F ( x) dx 
0 , and we find
1 1
3 0
c
1 c2  F ( x ) dx

which by Eq. (v), gives
 1 1 
φ ( x )  F ( x )  l  c2 
 3 0
F ( x ) dx   3 xc2 l


2 1
3 0
or φ( x)  F ( x)  F ( x ) dx  B(1  3 x ) (xiv)
where, B
2c2  2c1
1
Finally, if l = –2 and 0 (1  3x)F ( x) 
0, Eq. (xiv) provides infinitely
many solutions.
EXERCISE 4.1
Solve the following integral equations.

1
e x  l  2e x et  u(t ) dt
1. u( x )  0
p /2
2. φ ( x )  2 x  p  4  sin 2 x φ (x ) dx
0
1
s  l  st (s  t ) g(t ) dt
3. g(s)  0
1
4. f ( x ) (1  x )  1 ( xt  x t ) f (t ) dt
2 2 2
p
cos  l  sin( x  t ) φ (t ) dt
5. φ ( x )  0
1
f (s)  l  st (1  st ) g(t ) dt . Find its resolvent kernel also.
6. g(s)  1
7. Invert (or solve) the integral equation:

2p
(s) f (s)  l 
g (sin s cos t ) g(t ) dt
0
1
x  l  ( xt 2  x 2 t ) g(t ) dt
8. Solve g(s)  0
p /2
1  l  cos( x  t ) g(t ) dt , and find its eigenvalues.
9. Solve g( x )  0
Answers
ex

1. u( x ) , l  1/(e2  1)
1  l (e2  1)
2 2
2. f ( x )  2 x  p  (p sin x )/(p  1)
2 2
3. g(s)  [(240  60 l )  80 l s ]/(240  120 l  l )
4. f ( x ) 1  6 x  (25 / 9) x 2 or f ( x ) (1  x )2  4 x  (16 / 9) x 2
{4 cos x  2pl sin x}/(4  p 2 l 2 )

5. φ ( x ) 
1 
 3st 5s 2 t 2 
f (s)  l  
6. g(s)    f (t ) dt
1 3  2 l 5  2 l 

3st 5s 2 t 2
resolvent kernel R(s
, t; l ) 
3  2l 5  2l
7. g(s) = {2/(2 – l)} sin s, l ≠ 2
(240  60 l ) x  80 l x 2
8. g( x ) 
240  120 l  l 2
l (cos x  sin x )
9. g( x ) 
1 ,l
4/(p  2)
1  p (p  2)/4

Chapter 5
Integral Equations with

Symmetric Kernels
5.1 INTRODUCTION
In the previous chapters, Fredholm integral equations of the second kind have
been considered for any given kernal K(x, t) by having the eigenvalues and
corresponding eigenfunctions. In this chapter, the same equation is the main
motive, but now, the kernel is symmetric.
5.2 SYMMETRIC KERNAL

A kernel K(x, t) is said to be symmetric (also complex symmetric or Hermitian)
if
K ( x, t )  K (t , x ) (5.1)
where, the bar denotes the complex conjugate. If the kernel is real, the
symmetry reduces to equality.
K ( x, t )  K (t , x ) (5.2)
Theorem: If a kernel is symmetric, then all its iterated kernels are also
symmetric.
Proof: Let the kernel K(x, t) be symmetric. Then, by defintion,

K ( x, t )  K (t , x ) (5.3)
By definition, the iterated kernels K n ( x, t ), n I  are defined as
K1 ( x, t )  K ( x, t ) [5.4(a)]
b
K n ( x, t )   K ( x, z ). K n 1 ( z, t ) dz, n = 2,3, ... [5.4(b)]
a
b
Also, K n ( x, t )   K n 1 ( x, z ). K ( z, t ) dz , n = 2, 3, ... [5.4(c)]
a
56
Integral Equations with Symmetric Kernels 57
We shall use the principle of mathematical induction to prove the required
results.
By Eq. [(5.4(b)],
b
K 2 ( x, t )   K ( x, z ). K1 ( z, t ) dz
a
Now, by Eq. [(5.4(a)],

b
K 2 ( x, t ) = ∫ K ( x, z ). K ( z , t ) dz
a
and By Eq. (5.3)

b
K 2 ( x, t )   K ( z, x ). K (t , z ) dz
a
b
K 2 ( x, t )   K (t , z ). K ( z, x ) dz
a
By Eq. [(5.4(a)],
b


K 2 ( x, t ) a K (t, z)  K1 (z, x) dz
Thus, K 2 ( x, t )  K 2 (t , x )
which shows that K2(x, t) is symmetric by definition, and the required result
is true for n = 1, 2.
Let Kn(x, t) be symmetric for n = m. Then, by definition,
K m ( x, t )  K m (t , x ) [5.4(d)]
We shall show that Kn(x, t) is also symmetric for n = m + 1, i.e.,
K m 1 ( x, t )  K m 1 (t , x ) [5.4(e)]
Now, by Eq. [5.4(b)],
b
K m
1 ( x , t ) a K ( x, z)  K m (z, t ) dz
Now, using Eqs. (5.3) and [5.4(d)],
b

K m
1 ( x , t ) a K (z, x)  K m (t, z) dz
b

K
m 1 ( x , t ) a K m (t, z)  K (z, x) dz
which upon using Eq. [5.4(c)] becomes
K m 1 ( x, t )  K m 1 (t , x ) , which is the R.H.S. of Eq. [5.4(e)].
Thus, the iterated kernel Kn(x, t) is symmetric for n = 1, 2; and it is also
symmetric for n = m + 1 whenever it is true for n = m. Hence, by mathematical
induction, Kn(x, t) is symmetric for n = 1, 2, ... .
5.3 REGULARITY CONDITION

In our study, the functions are either continuous or integrable or square
integrable. When an integral sign is used, it is to be taken as Lebesgue integral.

Moreover, we know that a function which is Riemann integrable, it is also
Lebesgue integrable. By a square integrable function f(x), we mean that
b
a | f ( x) |
2
dx  
p /2
For example, 0 sec 2 x dx   and it (sec x) is not square integrable.
A square integrable function f(x) is also called L2 function. A kernel K(x, t)
is an L2 function if it satisfies the following three conditions:
b b
  | K ( x, t ) | dx dt   ,
2
1. V x [ a, b], V t [ a, b]
a a
b
 | K ( x, t ) | dt   , V x [a, b], and
2. 2
a
b
 | K ( x, t ) | dx   ,
2
3. V t [ a, b]
a
5.4 INNER OR SCALAR PRODUCT OF TWO FUNCTIONS

The inner or scalar product of two complex functions f and y of real variable
b
(f ,y )  f ( x ) y ( x ) dx ,
x, a  x  b is denoted by (f, y) and is defined as 
a
where the bar denotes the complex conjugate.
The norm of a function f(x) is given by the following relation:
1/2 1/2
|| f ( x ) ||  f ( x ).f ( x ) dx   b | (f ) x |2 dx 
b



 a   a 
A function f(x) is said to be normalised if ||f(x)|| = 1. It follows that a non-

null function (whose norm is not zero) can always be normalised by dividing
it by its norm.
Note: For the Fredholm linear operator K,
b

Kf a K ( x, t ) f (t ) dt
The operator adjoint to K is
b

Ky a K (t, x) y (t ) dt
The two operators Kf and Kψ are connected as follows:
(Kf , y )  (f , K y )
For a symmetric kernel, it reduces to (Kf , y )  (f , K y ) , which means that a
symmetric kernel is self-adjoint. Now, since the permutation of factors in a scalar
product is equivalent to taking the complex conjugate, so (f , K f )  (Kf , f ).
Combining this with (Kf ,y )  (f , Ky ) , we find that for a symmetric kernel,
the inner product (Kf,f) is always real. The converse of this is also true.
5.5 ORTHOGONAL SYSTEM OF FUNCTIONS

A finite or an infinite set {fk(x)} defined on an interval a  x  b is said
b
to be an orthogonal set if (fi,fj) = 0 or
2p
a fi ( x).f j ( x) 
dx 0, i  j . For
instance, 0 sin x.cos x dx  0 ; so sin x and cos x are orthogonal functions for
0  x  2p .
If none of the elements of this set is a zero vector, then it is called proper
orthogonal set. The set {fi(x)} is orthonormal if
b 0, i  j
(fi , f j ) f ( x ) f j ( x )dx
a i

1, i  j

Any function f(x) for which ||f(x)|| = 1 is said to be normalised.
Given a finite or an infinite (denumerable) independent set of functions
{y1, y2, ..., yk ...} we can construct an orthonormal set {f1, f2, ..., fk ...}.
We have an important theorem (without proof) relating such functions.
Riesz–Fisher Theorem: If {fi(x)} is a given orthogonal system of functions
in L2 and {ai} is a given sequence of complex numbers such that the series

 | a i |2 converges, then there exists a unique function f(x) for which ai are
i 1
the Fourier coefficients with respect to the orthonormal system {fi(x)} and to
which the Fourier series converges in the mean, i.e.,

|| f ( x )   a i .fi ( x ) ||  0 as n  ¥
i 1
5.6 FUNDAMENTAL PROPERTIES OF EIGENVALUES AND

EIGENFUNCTIONS OF SYMMETRIC KERNELS
We consider the following symmetric integral equation:
b
l  K ( x, t ). g(t ) dt  f ( x ) or lKg = f ; K ( x, t )  K (t , x )
a
Property 1: The eigenvalues of a symmetric kernel are real.

Proof: Let l and f(x) be an eigenvalue and a corresponding eigenfunction
of kernal K(x, t). Then, by definition of eigenfunction,
b
f ( x )  l  K ( x, t ) f (t ) dt
a
or, f ( x )  l Kf ( x )  0
Multiplying by f ( x ) and integrating with respect to x from a to b, we obtain
|| f ( x ) ||2  l (Kf , f ) 0
or l  || f ( x ) ||2 /(Kf , f )
r r
Since both N and D for R.H.S. are real, hence l is real.
Property 2: The eigenfunctions of symmetric kernel, corresponding to

different eigenvalues are orthogonal.
Proof: Let f1 and f2 be eigenfunctions corresponing to eigenvalues l1 and
l2, respectively, where l1 ≠ l2. Then by definition, we have
f1  l1Kf1 
0 (5.5)
f2  l2 K f2 
0 (5.6)
Since l2 is real, Eq. (5.6) may written as
f2  l2 K f2 
0 (5.7)
Multiplying Eq. (5.5) by l2 f2 and Eq. (5.7) by l1f1, subtracting and
integrating, we get  recollect (f ,y )   f ( x ) y ( x ) dx  ,
b
 a 
l1l2    f2 ( x ) K ( x, t ) f1 (t ) dt dx
b b
(l2  l1 )(f1 , f2 ) 
 a a

b b  (5.8)
f ( x ) K ( x, t ) f2 (t ) dt dx 
a a 1 
Since kernel K(x, t) is symmetric, we have
K ( x, t )  K (t , x ) (5.9)
Using Eq. (5.9), we find that R.H.S. of Eq. (5.8) vanishes, and so, we get
(l2  l1 )(f1 , f2 ) 
0 (5.10)
Now, since l2 ≠ l1, Eq. (5.10) reduces to (f1,f1) = 0,
which means f1 and f2 are orthogonal.
Property 3: The multiplicity of any non-zero eigenvalue is finite for every
b b
a a | K ( x, t ) | dx dt is finite.
2
symmetric kernel for which
Property 4: The eigenvalues of a symmetric L2 kernel forms a finite or an
infinite sequence {ln} with no finite limit point.
Property 5: The set of eigenvalues of the second iterated kernel coincide
with the set of squares of the eigenvalues of the given kernel.
Property 6: The sequence of eigenfunctions of a symmetric kernel can be
made orthonormal.
5.7 HILBERT–SCHMIDT THEOREM

If f(x) can be written in the form
b
f ( x )   K ( x, t ) h (t ) dt (5.11)
a
where K(x, t) is a symmetric–kernel and h(t) is an L2–function, then f(x) can
be expanded in an absolutely and uniformly convergent Fourier series with
respect to the orthonormal system of eigenfunctions f1 ( x ), f2 ( x ),...fn ( x ),... of
the kernel K(x, t).

f ( x )   f nfn ( x ) (5.12)
where, n 1
f n  ( f , fn ) (5.13)
The Fourier coefficients fn of the functions f(x) are related to the Fourier
coefficients hn of the functions h(x) by the following relations:
fn = hn/ln (5.14)
and hn = (h, fn) (5.15)
where ln are the eigenvalues of kernel K(x, t).
Proof: Let K(x, t) be a non-null, symmetric kernel which has a finite or
an infinite number of (real and non-zero) eigenvalues, ordering them in the
sequence
l1, l2, ..., ln,... (5.16)
in such a way that each eigenvalue is repeated as many times as its multiplicity.
We further agree to denumerate these eigenvalues in the order that corresponds
to their absolute value, i.e.,
0  | l1 |  | l2 |   | ln |  | ln 1 | 
Let f1 ( x ), f2 ( x ), , fn ( x ) (5.17)
be the sequence of eigenfunctions corresponding to the eigenvalues given by
the sequence shown in Eq. (5.16) and arranged in such a way that they are
no longer repeated and are linearly independent in each group corresponding
to the same eigenvalue.
Thus, to each eigenvalue lk in Eq. (5.16), there corresponds just one
eigenfunction fk(x) in Eq. (5.17). Further, we suppose eigenfunctions fk(x) in
Eq. (5.12) have been orthonormalised.
Now, the Fourier coefficients fn of the function f(x) with respect to the
orthonormal system {fn(x)} are [given by Eq. (5.13)]
f n (
f , fn ) (Kh
, fn ) (h, K fn )
1 h
fn  (hfn ) n , using the self-adjoint property of the operator and the
ln ln
relation lnKfn = fn.
Hence, the Fourier series for f(x) is given by Eq. (5.12) and is expressed as

hn
l n f
f ( x) ~
( x )   fnfn ( x) (5.18)
n 1n n 1
We now estimate the remainder term of the series Eq. (5.18), as shown
below:
2
n p
fk ( x )  h  p 2   h  p | fk ( x ) |2 
 k lh    hk   
 k   k lk2 

k
n 1 k n 1 n 1

 h p    f 2 ( x) 
   hk2    k 2  (5.19)
 k
n 1 1 lk 
  k
Using Bessel’s inequality,

|fn ( x ) |2 b
 ln2
  | K ( x t ) |2 dt  C12
a
n1
We find that the above series is bounded. Moreover, since h(x) is an

L2 function, it follows that the series  hk2 is convergent and the partial
nP k 1
sum  hk2 can be made arbitrarily small. Hence, Eq. (5.18) converges
k n  1
absolutely and uniformly.
We now proceed to show that Eq. (5.18) converges to f(x) in the mean.
For this purpose, let us denote its partial sum as
n
h
y n ( x )   m fm ( x ) (5.20)
m  1 lm
and estimate the value of || f ( x )  y n ( x ) || .

n
hm
Now, f ( x)  y n ( x) 
Kh   f ( x ) [For hm, refer to Eq. (5.15)]
lm m
m 1
n
(h, fm )
Kh  
 K ( m 1) h (5.21)
fm ( x ) 
m 1 l m
where K(n+1) is the truncated kernel. From Eq. (5.21), we have
( n 1) 2
2
|| f ( x )  y n ( x ) || || K h || (K ( n 1) h, K ( n 1) h)
h, K ( n 1) K ( n 1) h) (h, K 2( n 1) h) (5.22)

 (

where we have used the self-adjointness property of kernel K(n+1) and also the
relation K ( n  1) K ( n  1) h  K 2( n  1) .
We know that the set of eigenvalues of the second iterated kernel coincides
with the set of squares of the eigenvalues of the given kernel. Using this
property, we see that the least eigenvalue of kernel K 2( n  1) is equal to
1  (h, K 2( n  1) h) 
ln 2  1 . Again,  max   (5.23)
ln21  (h, h) 
where we have omitted the modulus sign from the scalar product (h,K2(n+1)h)
because it is a positive quantity.
Combining Eq. (5.22) and (5.23), we find
(h, h)

2
|| f ( x )  y n ( x ) || (h, K 2( n  1) h)  (5.24)
ln21
Since ln+1  ¥, Eq. (5.24) gives
|| f ( x )  y n ( x ) ||  0 as n  ¥ (5.25)
Now, we use the relation
|| f ( x )  y ( x ) ||  || f ( x )  y n ( x ) ||  || y n ( x )  y ( x ) || (5.26)
where y(x) is the limit of the series with partial sum yn.
As shown above, the first term on the R.H.S. of (5.26) tends to zero, and to
show that the second term of R.H.S. of (5.26) also tends to zero, we proceed
as follows:
Since Eq. (5.18) converges uniformly, we have, for an arbitrarily small
and positive quantity , | y n ( x )  y ( x ) | , when n is sufficiently large.
|| y n ( x )  y ( x ) ||  (b  a)1/2
and hence || y n ( x )  y ( x ) ||  0
Hence, Eq. (5.26) shows that f(x) = y(x), and thus, the result follows.
5.8 SCHMIDT’S SOLUTION OF NON-HOMOGENEOUS

FREDHOLM INTEGRAL EQUATION OF THE SECOND KIND
Consider
b
( x) f ( x)  l 
y K ( x, t ) y (t ) dt (5.27)
a
where K(x, t) is continuous, real and symmetric kernel and l is not an

eigenvalue.
Statement of Hilbert–Schmidt theorem:
Let F(x) be generated from a continuous function y(x) by the operator
b
l  K ( x, t )  y(t )dt,
a
where K(x, t) is continuous, real and symmetric,
b
so that F ( x ) l  K ( x, t )  y(t )dt (5.28)

a
The function F(x) can be expressed over the interval (a, b) by a linear
combination of the normalised eigenfunctions of homogeneous integral equation
b
y( x ) l  K ( x, t )  y(t )dt (5.29)
a
having K(x, t) as its kernel.
Procedure of solution
From Eq. (5.27), we have
b
( x ) l  K ( x, t )  y(t )dt (5.30)
y( x )  f
a
Since this equation is of the form Eq. (5.28), it follows from Hilbert–Schmidt
theorem
 am fm ( x) , a  x  b (5.31)
y( x )  f ( x ) 
m 1
where fm(x) (m = 1, 2, 3,...) are the normalised eigenfunctions of homogeneous

integral Eq. (5.29).
Let lm be the corresponding eigenvalues of Eq. (5.29), where l ≠ lm for
m = 1, 2, 3,...
Since fm(x) is normalised, we have
b 0, m  n
a fm ( x) .fn ( x) dx  1, m  n (5.32)
Multiplying both sides of Eq. (5.31) by fm(x) and then integrating with
respect to x from a to b, we get
b b
a y( x).fm ( x) dx  a f ( x).fm ( x) dx
b b
= a1  f1 ( x ) fm ( x )dx   am  fm ( x ) fm ( x ). dx  . (5.33)
a a
b
Now, let Cm   y( x ) fm ( x ) dx (5.34)
a
b
and f m   f ( x ) fm ( x ) dx (5.35)
a
Then, by using Eq. (5.32), Eq. (5.33) provides

Cm − f m = 0 + 0 + + am + 0 (5.36)
And now, multiplying both sides of Eq. (5.27) by fm(x) and then integrating
with respect to x from a to b, we get
b

a
b b
 b
y( x ) fm ( x ) dx  f ( x ) fm ( x ) dx  l   K ( x, t ).y(t )dt
a a a  fm ( x ) dx
Now, using Eqs. (5.34) and (5.35) and interchanging the order of integration
C
m fm  l 
b
a
y(t ) a
b

K (t , x ) fm ( x )dx dt (5.37)
Here, we have used the symmetric property of K(x, t).

Further, since fm(x) is eigenfunction corresponding to the eigenvalue lm of
Eq. (5.29), by definition, we have
b
fm ( x ) lm 
 K ( x, t ) fm (t ) dt
a
[Since eigenfunction has to satisfy the integral equation]
b
fm ( x ) lm  K ( x, z ) fm ( z ) dz

a
[By changing the variable of integration (t to z)]

b
fm (t ) lm  K (t , z ) fm ( z ) dz

a
[By changing the argument x]
b
fm (t ) lm  K (t , x )  fm ( x ) dx

a
[By again changing the variable of integration]

b fm (t )
or a K (t, x).fm ( x) dx  lm
(5.38)
Then, Eq. (5.37) gives
b fm (t ) b 1
Cm f m  l  y(t )  dt f m  l   y( x ) fm ( x ) dx
a lm a l
m

which upon using Eq. (5.34) shapes as
l Cm
C
m fm  (5.39)
lm
Now, by Eq. (5.36), we have [by eliminating Cm (just by putting Cm)]
l
am  f m  f m  (am  f m )
lm
l
giving am  f m (5.40)
lm  l
Upon substituting this values of am in Eq. (5.31), the required solution of
Eq. (5.27) is
 l 
  l  l fm  fm ( x)
y( x )  f ( x ) 
m 1 m
fm
or y( x
) f ( x)  l S f ( x ) (5.41)
lm  l m
Further, using Eq. (5.35) for fm, we get (variable of integration changed)
fm ( x ) b
( x) f ( x)  l 
y a f (t ) fm (t ) dt
m lm  l
 f ( x )fm (t ) 
b
or ( x ) f ( x )  l   m
y  f (t ) dt
 m lm  l 
a
b
or ( x ) f ( x )  l  R ( x, t; l ) f (t ) dt (5.42)
y
a
where the resolvent kernel R(x, t; l) is given by

fm ( x ).fm (t )
R ( x, t ; l )   (5.43)
m lm  l
We analyse the following three cases:
Case 1: Unique solution: l ≠ lm, am is given by Eq. (5.40) to be substituted

in Eq. (5.31) to get the unique solution which is given by
fm
y( x
) f ( x)  l S f ( x ) (5.44)
lm  l m
Case 2: No solution: Let the kth eigenvalue be lk = l. Also let

b
fk a f ( x) fk ( x) dx  0 ,
which means that fk(x) is not orthogonal to f(x).
f f ( x)
Now, because of the presence of the term k m in the solution, this term
is not defined; so, no solution is obtained. lk  l
Case 3: Infinitely many solutions: Let l = lk, and fk = 0. Then by Eq. (5.39)
lCm C  0  l C Ck  Ck
C fm  , k k
m
lm l
This being an identity does not impose any restriction on Ck, with the result ak
[refer to Eq. (5.40)] is of form zero/zero and becomes arbitrary. In this situation,
we express Eq. (5.41) as follows:
lm
y( x ) f ( x )  A fk ( x )  l ' fm ( x ) (5.45)
m lm  l
where primed S implies that we shall omit m = k in the summation and A is
an arbitrary constant. The solution given in Eq. (5.45) due to arbitrary nature
of A shows that given Eq. (5.27) possesses infinitely many solutions.
EXAMPLE 5.1: By using Hilbert–Schmidt theorem, solve the following

symmetric integral equation:
1
y( x ) ( x  1)2   ( xt  x 2 t 2 ) y(t )dt
1
1
Solution: The given equation is y( x ) = ( x + 1)2 + ∫ ( xt + x 2 t 2 ) y(t )dt (i)
−1
Comparing Eq. (i) with Eq. (5.27), i.e.,
b
) f ( x)  l  
y( x K ( x, t ).y(t )dt (ii)
a
( x 1)2 , l 
we get f ( x )  ( xt x 2 t 2 ) (iii)
1, K ( x, t ) 
First, we determine the eigenvalues and the corresponding normalised
eigenfunctions of Eq. (i) after deleting its non-homogeneous part. So, let
1
) l
y( x ( xt  x 2 t 2 )  y(t )dt (iv)
1
1 1
y( x ) l x  t .y(t ) dt  l x 2  t 2 y(t ) dt (v)
1 1
1
Let 1 t y(t )dt  c1 (vi)
1
1 t
2
and .y (t ) dt  c2 (vii)
y ( x ) l c1 . x  l c2 . x 2 (viii)
So, Eq. (v) reduces to 
and y (t ) l c1 . t  l c2 . t 2 (ix)

1
1 t (l c1 . t  l c2 . t
2
Then, by Eq. (vi) c1 ) dt
1 1
 t3   t4 
or c1 l c1    l c2  
 3  1  4  1
2
c1 l c1  0.c2
3
 2l 
or  1   c1  0.c2  0 (x)
 3 
Similarly, by Eq. (ix) and (vii)
1

c2 1 t 2 (l c1 .t  l c2 t 2 )dt
1 1
 t4   t5 
or c2 c1l    c2 l  
 4  1  5  1
 2l 
or 0. c1  1   c2 
0 (xi)
 5 
Equation (x) and (xi) will provide non-trivial values of c1 and c2 only if
2λ
1− 0
3
D( λ ) = = 0 (xii)
2λ
0 1−
5
Giving l = 3/2, 5/2 as the required eigenvalues.
Determination of eigenfunction corresponding to l1 = 3/2
3
Putting 
l ( l1) in Eq. (x) and (xi), we get
2
 2 3
0 and 0. c1  1  .  c2 
0. c1  0. c2  0
 5 2
we find that c2 = 0 and c1 is arbitrary.
3
Then, by Eq. (viii), one eigenfunction is y1 ( x )  c1 x or y1(x) = x by taking
2
3c1
1
2
Now, the corresponding normalised eigenfunction f1(x) is given by
y1 ( x ) x x 6
f1 ( x )   (xiii)
1 1 2
 1 {y ( x )}2 dx  2  1 x 2 dx  2
 1 1   1 
5
Determination of eigenfunction corresponding to l 2 
2
5
Putting 
l ( l
2 ) in Eqs. (x) and (xi), we get
2
 2 5
0. c1  1  .  c2  0 0. c1  0. c2 0
 5 2
and
 2 5  5
1  .  c1  0. c2  0 1   c1  0. c2 0
3 2 3
We find that here, c1 = 0 and c2 is arbitrary.
Now, substituting these c1 and c2 in Eq. (viii), the eigenfunction
5
corresponding to l2  is
2
5
y2 ( x ) 0  c2 x 2
2
2 5
y2(x) = x , (by taking c2  1 )
2
Further, the corresponding normalised eigenfunction f2(x) is given by
y2 ( x ) x2 10 2
 f2 ( x )  1
 1
x
2 (xiv)
 1 y ( x ) 2 dx  2  1 x 4 dx  2
 1  2    1 
Now, by Eqs (iii) and (xiii) [also refer to Eq. (5.35)],
1 1  6x 

f1 1 f ( x ). f1 
( x )dx 1 ( x  1)2 
 2 
 dx
1
6 1 6  x 4 2 x3 x2  2 6 (xiv)
1
3 2

f1 ( x  2 x  x )dx
    
2 2  4 3 2 3
1
Similarly, by Eqs. (iii) and (xiv),
1 1 10 2
1 f ( x) f2 ( x
)dx  ( x  1) 2

f2 x dx
1 2
1
10  x5 2 x 4 x3  8

f2     10 (xv)
2  5 4 3  1 15
3

Also, from Eq. (iii) l = 1 and l1  , l2 5 ;
2 2
Thus, we see that l  l1  l2 , the integral equation, i.e., Eq. (i) will possess
a unique solution, which is given by
2
fm
y
( x) f ( x)  l  f ( x ) (xvi)
lm  l m
m 1
 f f f f 
y( x ) ( x  1)2  1  1 1  2 2 
 l1  l l2  l 
 2   x 6   8 10   10 2 
 6 x 
2
3   2   15   2 
y( x ) ( x  1)  
3 5
1 1
2 2
25 2
or y( x
) x  6 x  1 , is the solution of Eq. (i).
9
EXAMPLE 5.2: Using Hilbert–Schmidt theorem, find the solution of the
following symmetrical integral equation
3 1
y( x )  ( x 2  1)  1 ( xt  x
2 2
t )  y(t ) dt
2
Solution: The given equations is
3 1
2 1
y( x )  ( x 2  1)  ( xt  x 2 t 2 )  y(t )dt (i)

b
( x ) f ( x )  l  K ( x, t ). y (t ) dt (ii)
y
a
We have
3
f ( x)  ( x 2  1), l , K ( x, t ) xt  x 2 t 2 (iii)
2
First, we determine the eigenvalues and the corresponding normalised
eigenfunctions of Eq. (i) after deleting its non-homogeneous part. Then, it is for
1
y( x ) l  ( xt  x 2 t 2 )  y(t )dt (iv)
1
Equation (iv) is same as Eq. (iv) of Example 5.1. So, we take the eigenvalues
and the corresponding normalised eigenfunctions and these are
3 5 x 6 10 x 2

l1  , l2 , f
1 ( x) , f
2 ( x)
2 2 2 2
1 1 x 6
1 f ( x)f1 ( x)dx 
1 ( x  1) 2 dx 
2
Now, f1 ( x )  0 (v)
1
and f2 ( x )   f ( x )f2 ( x )dx
1
1
1 10 x 2 10  x 5 x 3  8 10
f2 ( x )  1 ( x 2  1)
2
dx  2    
2  5 3  15
(vi)
0
3
Here, we find that l  l1 and l ≠ l2, so here, infinitely many
2
solutions will exist [case 3], and then the solution is given by
2
fm
y( x ) f ( x )  Af1 ( x )  l ' .f ( x ) (vii)
m 1 lm  l m
Here, S' means that the term for m = 1 must be neglected. So, we get
f2
y( x ) f ( x )  Af1 ( x )  l f ( x)
l2  l 2
(8 10)
 x 6  3 2
15 x 10
y( x )  ( x 2  1)  A   
 2  2 53 2
2 2
A 6
or y( x )  x 2  1  c1 x  4 x 2 , c1 
2
2
or y( x )  5 x  c1 x  1 is the required solution, c1 being an arbitrary constant.
EXAMPLE 5.3: Solve the following symmetric integral equation by Hilbert–

Schmidt theorem:
p
y( x ) 1  l  cos( x  t )  y (t ) dt
0

p
y( x ) 1  l  cos ( x  t )  y(t )dt (i)
0

b
( x ) f ( x )  l  K ( x, t ) y (t ) dt (ii)
y
a
We get f(x) = 1, l = l, K(x, t) = cos (x + t) (iii)

We first determine the eigenvalues and the corresponding normalised
eigenfunctions of Eq. (i) after deleting its non-homogeneous part, i.e., let
p
y( x ) l  cos( x  t )  y (t ) dt

0 (iv)
p
y( x ) l  (cos x cos t  sin x sin t )  y (t ) dt

or 0
p p
or y( x ) l cos x  cos t  y(t ) dt  l sin x  sin t  y(t ) dt (v)

0 0
p
Let c1 0 cos t  y(t ) dt (vi)
p
and c2 0 sin t  y(t ) dt (vii)
Then, Eq. (v) becomes

y( x ) c1l cos x  c2 l sin x (viii)
and
y(t ) c1l cos t  c2 l sin t
(ix)
p
Then Eq. (vi) gives c1  cos t [c1l cos t  c2 l sin t ] dt
0
c1l p c l p

c1  
2 0
(1  cos 2t ) dt  2  sin 2 t dt
2 0
p p
c1l  sin 2t  c2 l  cos 2t 
c1  t  
2  2 0 2  2 0

c1l

c1 p
2
c1  (2  l p )  0  c2  (x)
0
Similarly, using Eq. (ix), (vii) gives
p

c2  sin t c1l cos t  c2 l sin t  dt
0
c1l p c l p


c2 
2 0
sin 2t dt  2  (1  cos 2t ) dt
2 0
p p
c1l  cos 2t  lc  sin 2t 
or c2    2 t  2 
2  2 0 2  o
c l
0  2 (p  0 )
c2 
2
or 0  c1  (2  l p ) c2 
0 (xi)
Equations (x) and (xi) will have a non-trivial solution if
2  lp 0
D( l )  0
0 2  lp

 (2  lp )(2  lp ) 
0
2 2 2 2
l , 
,(say) l1  , l2
p p p p
So, l1 and l2 are the eigenvalues.
Determination of eigenfunction corresponding to

2
For this value of 
l l
1 , Eqs. (x) and (xi) give
p
c1.0 + 0.c2 = 0 and 0.c1 + 4c2 = 0
This means c2 = 0 and c1 is arbitrary.
2
Putting these values in Eq. (viii) and recollecting that l1  , we have the
eigenfunction y1(x) given by p
2 2c1
y1 ( x ) 
 c1  cos x  cos x , by taking 1
p p
The corresponding normalised eigenfunction f1(x) is given by
y1 ( x ) cos x cos x
f1 ( x )  1/2

 p p 1  cos 2 x
{y1 ( x )}2 dx 
p
 cos x dx
2

 0  0 0 2
dx
1/2
cos x cos x  2 

f1 ( x )     cos x (xii)
1 sin 2 x 
p p / 2 p 
 x  
2 2 0
–2
Determination of eigenfunction corresponding to l 2 =
p
2
For this value of 
l l2 , Eqs. (x) and (xi) give
p
4c1 + 0.c1 = 0 and 0.c1 + 0.c2 = 0
This means c1 = 0 and c2 is arbitrary.
2
Putting these values in Eq. (viii) and recollecting that l  , we have
the eigenfunction p
y2 ( x )  0  ( 2 / p )  c2 sin x
 2c 
or y2(x) = sin x,  Taking 2  1
 p 
Then, the corresponding normalised eigenfunction f2(x) is given by
y2 ( x ) sin x 2

f2 ( x )   sin x (xiii)
 p y ( x ) 2 dx 
1/2 p p

 0 2
 
 sin x dx
0
2
p p 2
Now, f1 ( x ) 0
f ( x ).f1 ( x ) dx 
0
1.
p
cos x dx 0 (xiv)
p p 2 2
and  f ( x) f2 ( x) dx 
f2 ( x ) 
0  1 0 p
sin x dx 
2
p
(xv)
Case 1: Let l ≠ l1 and l ≠ l2, then Eq. (i) will possess unique solution
given by
2
fm
( x) f ( x)  l 
y fm ( x )
 l m l
m 1
f f
1  l . 1 f1 ( x )  l . 2 f2 ( x )
y( x ) 
l1  l l2  l
Now, substituting the values of l1, f1, f1 and l2, f2, f2 , we get
1/2 1/2
l  2  2
y( x ) 1  0  2  
l
 
p 
sin x since f1  0 
 2 
 p   l
4l sin x
or y( x )  1  (xvi)
2  pl
2
Case 2: Let l  l2   . Since f2 ≠ 0, so Eq. (i) possesses no solution.
p
[Note: f1 = 0, and denominator of last term becomes zero.]
Case 3: Let l = l1 = 2/p. Since f1 = 0, there exists infinitely many solutions
given by
2
fm
y( x ) f ( x )  Af1 ( x )  l  ' fm ( x ) (xvii)
m 1 l m  l
where dash indicates that in the sum, the term corresponding to m = 1 is to

be omitted. Thus Eq. (xvii) reduces to
l f2
y( x ) f ( x )  A f1 ( x )  f ( x)
l2  l 2
1/2 1/2 1/2
 2 2/p  2  2
y( x ) 
1 A   cos x  2   .  sin x
p  2 2 p  p 
 
p p
2 sin x
y( x ) 
1  c cos x 
p
2A
where, c  is an arbitrary constant.
p
EXAMPLE 5.4: Using Hilbert–Schmidt method, solve
1
x  l  K ( x, t )  y(t ) dt
y( x ) 
0
 x (t  1), 0 x t
where, K ( x, t )  
 t ( x  1), t  x 1

1
x  l  K ( x, t )  y(t ) dt (i)
y( x ) 
0
 x (t − 1), 0 ≤ x ≤ t
where K = (ii)
 t ( x − 1), t ≤ x ≤1
Comparing Eq. (i) with Eq. (5.27), we get
f(x) = x (iii)
First, we begin with the determination of eigenvalues and the corresponding
eigenfunctions of
1
y( x ) l  K ( x, t )  y(t ) dt (iv)
0
( x ) l   K ( x, t )  y (t ) dt   K ( x, t )  y (t ) dt 
x 1
or y
 0 x 
Now, from Eq. (ii),
x 1
y( x ) l  t ( x  1)  y(t ) dt  l  x (t  1)  y (t ) dt (v)

0 x
Recollect Leibnitz’s rule of differentiation under the sign of integration,

which is Eq. (1.20).
[Note the procedure when the kernel is defined in this manner]
d H ( x) H ( x ) F dH dG


dx G ( x ) 
F ( x, x ) dx 
G ( x ) x
dx  F [ x, H ( x )]
dx
 F[ x, G( x )]
dx
x  d d
y
'( x ) 0 x {lt ( x  1). y (t )}dt  l x ( x  1). y( x). dx ( x)  l.0 ( x  1).y(0) dx (0)
1  d d
 {l x (t  1). y (t )}dt  l x (1  1). y(1). (1)  l .x ( x  1).y( x ) ( x )
x x dx dx
x
or y '( x )  l t . y (t ) dt  l x ( x  1).y ( x ).1  0

0

1
  l (t  1).y(t ) dt  0  l x ( x  1).y ( x ).1
x
x 1
or y
( x ) 0 lt .y (t ) dt  x l (t  1). y(t ) dt (vi)
Again, differentiating Eq. (vi) with respect to x, we get
x  d d

y "( x ) 0 x [lt . y(t )] dt  l x. y ( x) dx ( x)  l .0. y (0) dx (0)
1  d d
 [ l (t  1).y (t ) ]dt  l (1  1). y (1) (1)  l ( x  1) y( x ). ( x )
x x dx dx
or y "( x )  0  l x. y( x )  0  0  0  l ( x  1). y( x )
or y "( x )  l .y ( x ) 
0 (vii)
by Eq. (v) y(0) = 0 (viii)
and y(1) = 0 (ix)
*
Equation (vii) with Eqs. (viii) and (ix) is Sturm–Liouville problem.
[refer to {r ( x ). y '}'  {q( x )  l p ( x )}y 
0 a x b
satisfying k1y + k2y¢ = 0 at x = a and l1y + l2y¢ = 0 at x = b]
We consider the different cases depending upon l
Case 1: Let l = 0, The solution of Eq. (vii) is
y(x) = Ax + B (x)
Using Eqs. (viii) and (ix), we get
B = 0, A + B = 0
 A = 0 = B, so for l = 0, the solution is y(x) = 0.
Thus, l = 0 is not an eigenvalue and y(x) = 0 is not an eigenfunction.
Case 2: Let l = m2 (m ≠ 0). Then, the solution of Eq. (vii) is
y Ae m x  Be  m x ;
Using Eq. (viii) and (ix), we get
A + B = 0, Ae m  Be  m 
0
giving A = 0 = B.
Thus y(x) = 0 is not eigenfunction.
* Refer to Chapter 8 of Differential Equations with applications published by RBD.]

Case 3: Let l = m2 (m ≠ 0). Then, the solution of Eq. (vii) is

y = A cosmx + B sin mx
Using Eq. (viii) and (ix), we get
0 = A and 0 = A cos m + B sin m
giving B sin m = 0.
In order that y(x) = 0 may not be again a similar trivial solution, we take
B ≠ 0, then sin m = 0 giving m = np, n  
m2 
\ l  n2 p 2
Thus, the required eigenvalues ln are
ln   n2 p 2, n = 1, 2, 3, ...
and
y( x )  B sin np x
Taking B = 1, we have the eigenfunction
yn ( x )  sin np x , n = 1, 2, 3, ... (xi)
The normalised eigenfunction fn(x) is given by
yn ( x ) sin np x
fn ( x )  1/2 1/2
 1 y ( x ) 2 dx   1 sin 2 np x dx 

 0 n    0 

or, fn(x) = 2  sin np x (xii)
1
Now, fn = 0 f ( x).fn ( x) dx ; f(x) = x [By Eq. (iii)]
1
fn = 0 x. 2 sin np x dx
   cos np x 1 1 cos np x 
fn = 2   x.   0  dx 
  np 0 np 
  cos np 1  sin np x  
1
fn = 2    
 np np  np  0 
 ( 1) n 1  ( 1) n 1 2
fn = 2  2 2 .0   (xiii)
 np n p  np
Now, two cases arise,
Case 1: Let l ≠ ln, n = 1,2,3,... i.e., l is not an eigenvalue. Then Eq. (i)
will possess a unique solution, given by

fn
( x) f ( x)  l 
y fn ( x )
n 1 l n  l

( 1) n 1 2 1
y(x) = x  l   2 2 2 sin np x
n 1 n p n p  l
2 l   1 sin np x
n
y(x) = x   (xiv)

p n 1 n n2 p 2  l 
Case 2: Let l = ln = –n2p2, n = 1, 2, 3, ...
Then since from Eq. (xiii), fn ≠ 0 for n = 1, 2, 3, ...
Hence, Eq. (i) will possess no solution.
PRACTICE QUESTIONS WITH INTERMEDIATE RESULTS

1. Solve the symmetric integral equation
b
( x ) f ( x )  l  K ( x ). K (t ). y(t )dt
y
a
b C .K ( x )

Here /  {K ( x )}2 dt , y1 ( x )
l1 1 b
a
a {K (t )} dt
2
b
= K(x) [By taking C /  {K (t )}2 dt  1 ]
a
K ( x)
so f1 ( x )  1/2
 b {K ( x )}2 dx 
 a 
1
b  b   b 2
f ( x ). f1 ( x )dx  a f ( x ). K ( x )dx  /  a {K ( x )} dx 
2
 f1 
a    
Case 1: Let l ≠ l1. Then, we have the unique solution given by

1
fm
( x) f ( x)  l 
y fm ( x )
m  1 lm  l

l
 f ( x) 
y(x) f1 .f1 ( x )
l1  l b
 f ( x) 
y(x)
l
.
a f ( x). K ( x)dx . K ( x)
1 1 1
 b {K ( x )}2 dx 
 a  l  b {K ( x )}2 dx 2  b {K ( x )}2 dx  2
 a  a

 
b
l . K ( x )  f ( x ). K ( x )dx
a
 f ( x) 
y(x) b
1  l  {K ( x )}2 dx
a
Case 2: Let l = l1 and assume f(x) is not orthogonal to f1(x), i.e.,

b
f1 a f ( x). f1 ( x) dx  0 ,
then we have no solution.
Case 3: Let l = l1 and f1 = 0, then, we express the solution as

y(x) = f(x) + Af1(x), (A is arbitrary)
1
 f ( x )  A K ( x )   {K ( x )}2 dx 
b 2
y(x)
 a 
 f ( x )  c. K ( x )
y(x)
1
where, c  A   {K ( x )}2 dx  is a constant, though arbitrary.

b 2
 a 
2. Determine the eigenvalues and the corresponding eigenfunction of the
following equation
2p
x  l  sin ( x  t ). y(t ) dt
y( x ) 
0
Also, find the solution when l is not an eigenvalue.

Hint: The given equation is
p
x  l  sin ( x  t ) .y(t ) dt (i)
y( x ) 
0
We first determine the eigenvalues and eigenfunction for

2p
y( x ) l  (sin x cos t  cos x .sin t ). y (t ) dt
0
y(x) = l c1 sin x  l c2 cos x

2p 2p
where, c1   cos t .y(t )dt and c2   sin t .y (t )dt
0 0
Clearly, 
y(t ) l c1 sin t  l c2 cos t
c
1 l p c2  c1  lp c
2 0
and c
2 l p c1  lp c1  c
2 0
1  lp 1 1
D (l )  0  l1  , l2 
lp  1 p p
1
Determination of eigenfunction for l1 
p
1
Taking 
l l 1 , we get c1 = c2 so that
p
c
y1 ( x )  1 (sin x  cos x )  (sin x  cos x )
p
c1
(By taking  1)
p
The corresponding normalised eigenfunction f1(x) is given by
1
y1 ( x ) /   {y1 ( x )}2 dx  2 (sin x  cos x )/ 2p

2p
f
1 ( x)  
 0
Determination of eigenfunction for λ 2 = − 1

π
1
Taking  l l 2 , we get c1 = –c2
p
so that y2(x) = sinx – cosx, (by taking c1/p = 1)
Thus, the corresponding normalised eigenfunction f2(x) is given by
1/2
y2 ( x ) /   {y2 ( x )}2 dx

2p
f
2 ( x)  (sin x  cos x )/ 2p
 0
2p 2p x (sin x  cos x )
Now, 0 f ( x) f1 ( x)dx 
f1  0 2p
 2p
dx 
2p 2p x (sin x  cos x )
and f2  0 f ( x ). f2 ( x )dx 
2p
0
dx   2p
In the question, it is given that l ≠ l1, l2; hence, Eq. (i) will possess
unique solution, which is given by
2
fm
y( x ) x  l l fm ( x )
m 1 m l

f .f ( x ) f .f ( x )
xl 1 1
y( x )  l 2 2
l1  l l2  l
y( x ) 
x
 
l  2p (sin x  cos x ) / 2p

l  2p (sin x  cos x )/ 2p  
1 1
l  l
p p
2l 2p 2 sin x 2lp cos x
or y( x ) 
x 
1  l 2p 2 1  l 2p 2
3. Solve the symmetric integral equation
1
) e x  l  K ( x, t ). y(t )dt
y( x
0
 sinh x .sinh (t  1)
 , 0 xt
 sinh 1
where, K ( x, t )  
 sinh t . sinh ( x  1) , t  x 1
 sinh 1
Hint: The given equation is

1
e x  l  K ( x, t )  y(t )dt (i)
y( x ) 
0
 sinh x .sinh (t − 1)
 , 0≤ x≤t [ii(a)]
 sinh 1
where, K ( x, t ) = 
 sinh t . sinh ( x − 1) , t ≤ x ≤ 1 [ii(b)]
 sinh 1
Here, f(x) = ex
First, we determine the eigenvalues and eigenfunction for the homogeneous
integral equation
1
y( x )  l  K ( x, t ). y(t )dt (iii)
0
x sinh t sinh( x  1) 1 sinh x sinh(t  1)

or y( x ) l
sinh1
t 0 
t x
. y(t )dt  l 
sinh1
. y (t ) dt (iv)
Following Leibnitz’s rule, differentiating Eq. (iv), we get
x sinh t cosh ( x  1) l sinh x sinh ( x  1) y( x )

 y '( x ) l  . y(t )dt  .1

0 sinh1 sinh1
1 cosh x sinh(t  1) l sinh x . sinh ( x  1)

l  . y(t )dt  y( x ).1

x sinh1 sinh1
Again differentiating,
x sinh t sinh ( x  1) l sinh x . cosh ( x  1)

 y ( x ) l  y(t )dt  y( x ).1

0 sinh1 sinh1
1 sinh x  sinh (t  1) l cosh x . sinh ( x  1)

l  y(t )dt  y( x ).1

x sinh1 sinh1
Using Eq. (iv), it reduces to
l y( x )
y "( x ) y( x ) 
sinh 1
sinh x cosh ( x  1)  cosh x .sinh ( x  1)

or y "( x ) y( x )  l . y ( x )  y " (1  l ) y 0 (v)
with y(0) = 0 [obtained by putting x = 0 in Eq. (iv)] (vi)
and y(1) = 0 [obtained by putting x = 1 in Eq. (iv)] (vii)
For the solution of Eq. (v), we consider the following three cases:
Case 1: Let 1  l 0l  1 , then by Eq. (v) y" = 0, giving y = Ax + B.
Now, by Eq. (vi), B = 0 and by Eq. (vii), A + B = 0, which means both
A = 0, B = 0,
Therefore, y = 0.
Thus it is not an eigenfunction, and correspondingly, l = –1 is not an
eigenvalue.
Case 2: Let l + 1 = m2(m ≠ 0). Then, the general solution of Eq. (v) is
y( x ) Ae m x  Be  µx .

For A and B, we use Eq. (vi) and Eq. (vii) giving
0 = A + B and0 Ae m  Be  m ;
and these equations provide A = 0 = B.
Thus, y(x) = 0 is not eigenfunction and l = m2 – 1 is not an eigenvalue.
Case 3: Let l + 1= – m2(m ≠ 0). Then Eq. (v) has its general solution as
y(x) = A cos mx + B sin mx
Using Eqs. (vi) and (vii), we get
A = 0, B sin m = 0
Clearly, B ≠ 0 (otherwise we reach at the previous cases)

\ m np , n  I (an integer)
\ 1  l m2  n2p 2 or l  1  n2p 2  

and so, the required eigenvalues are ln  1  n p  , n 
2 2
1, 2,3 ,....
Now, taking B = 1, the eigenfunctions are
yn(x) = sin npx, n = 1,2,3, ...
The normalised eigenfunction fn(x) is given by
yn ( x ) sin n p x

fn ( x )  1
 2 sin n p x (viii)
1
 sin n p xdx
2
 1{y ( x )}2 dx  2
 0
0

n
1 1 x
Then, fn 
0
f ( x ) . fn ( x )dx  e fn ( x ) dx
0
 
1
fn   e x . 2 sin np x dx
0
2 e x (sin np x  np cos np x ) 
1

fn 2 2  0
1 n p
2
fn  e np cos np  (  np )
1  n2p 2
np 2 
or fn 1) n  , n 1, 2,3 (ix)
1  e ( 
1  n2p 2 
Here, we consider two situations:
(a) Let l ≠ ln (means l is not an eigenvalue). Then Eq. (i) will possess
a unique solution given by

fn
( x) f ( x)  l 
y f ( x)
n 1 ln  l n


np 2 2 sin np x
y( x )  e x  l  2 2
[1  e( 1)n ]
n 1 1 n p (1  n2p 2 )  l

n[1  e( 1) n ]sin np x
or, ) e x  2pl 
y( x
n 1 [1  n2p 2 ][1  l  n2p 2 ]
(b) Let l ln 1  n2p 2 , n 1, 2,3 ,....

Then, from Eq. (ix), fn ≠ 0 for n = 1, 2, 3, ...
Hence, Eq. (i) will possess no solution.
4. Using Hilbert–Schmidt theorem, solve the following symmetric integral
equations:
1
x   ( x  1) .y (t )dt , l  l1 , l2
(a) y( x )  0

 (6l  12) x  4l 
 Ans : y( x )  2 
 l  12l  12 
1  x 3   6  4 3   ( x  t) y(t)dt
1
(b) y( x ) 
0
    3x 

[ Ans : y( x )  1  x 3  C 1  3  1   , C being an arbitrary constant.]
 2
EXERCISE 5.1
Using Hilbert–Schmidt theorem, solve the following symmetric integral
equations:
1 1
1. y( x )   x   y(t ) dt ,
2 0
1
2. y( x ) x  l 0 xs y(s)ds,
1
x  l  y(s)ds, l  1
3. y( x ) 
0
1
4. y( x )  (1  x 3)  (6  4 2) 0 ( x  t )  y(t )dt
cos p x  l  K ( x, x )  y(x )dx
5. y( x ) 
x ( x  1) 0  x  x
where, k ( x, x )  
Answers:  x (x  1) x  x  1
1
1. [y(x) = – x + c]
2
3. [y(x) = x + l/{2(1 – l)]
4. [ y( x )  (1  x 3)  (1  3 x / 2)  c(1  x 3) ]

Chapter 6
Solution of Integral
Equations of the Second Kind
by Successive Approximation
6.1 INTRODUCTION
In the previous chapters of the book, the solution of the integral equations
was mainly focused upon Fredholm integral equation. In this chapter the
development is for Volterra integral equation also. Moreover, where the
solution is not possible in closed form, successive approximate method is
also discussed. Apart from the various theorems, the chapter includes three
parts—(a) iterated kernels, (b) resolvent kernel, and (c) solution of integral
equations by applying resolvent kernel.
6.2 ITERATED KERNEL OR FUNCTION

1. Let us consider the following Fredholm integral equation of the second
kind:
b
f ( x ) l  K ( x, t )  y(t )dt (6.1)
y( x ) 
a
Then, the iterated kernels Kn (x, t), n = 1, 2, 3, ..., are defined as follows:
K1 ( x, t )  K ( x, t ) [6.2(a)]
b
14243
and  K ( x, z)  K n 1 (z, t )dz,

K n ( x, t ) 
a
n
2,3,...
b
[6.2(b)]
or K n ( x, t )  K n 1 ( x, z)  K (z, t )dz,
a
n2,3,...
2. Let the Volterra integral equation of the second kind be

x
f ( x ) l  K ( x, t )  y (t )dt (6.3)
y( x ) 
a
Then, the iterated kernels Kn(x, t), n = 1, 2, 3, ..., are defined as follows:

K1 ( x, t )  K ( x, t ) [6.4(a)]
84
Solution of Integral Equations of the Second Kind by Successive Approximation 85
x
14243
and
 K ( x, z)  K n 1 (z, t )dz,
K n ( x, t ) 
t
n
2,3, …
x
[6.4(b)]
or
 K n 1 ( x, z)  K (z, t )dz,
K n ( x, t ) 
t
2,3, …
n
6.3 RESOLVENT KERNEL OR RECIPROCAL KERNEL

We consider the solution of Fredholm integral equation of the second kind.
b
f ( x ) l  K ( x, t )  y (t )dt (6.5)
y( x ) 
a
and let it take the following form
b
f ( x )  l  R( x, t; l )  f (t )dt [6.6(a)]
y( x ) 
a
b
or y( x )  f ( x )  l  ( x, t; l )  f (t )dt [6.6(b)]
a
Here, R(x, t; l) or (x, t; l) is known as resolvent kernel of Eq. (6.5).

Analogously, we have the resolvent kernel for Volterra integral Eq. (6.3).
We consider the following theorem without proof:
Theorem: The mth iterated kernel Km(x, t) satisfies the relation
b


K m ( x, t ) a Kr ( x, y)  K m r ( y, t ) dy
where, r is any positive integer less than m.
6.4 SOLUTION OF FREDHOLM INTEGRAL EQUATION OF THE

SECOND KIND BY SUCCESSIVE SUBSTITUTION
b
f ( x ) l  K ( x, t )  y(t )dt (6.7)
Theorem: Let y( x ) 
a
be the given Fredholm integral equation of the second kind. Let

1. K(x, t) / 0 be real and continuous in the rectangle R, for which a  x  b,
a  t  b. Also, assume |K(x, t)|  M in R.
2. f(x) ≠ 0 be real and continuous in the interval I for which a  x  b.
Also, assume | f(x)|  N in I.
3. l be a constant such that |l | < 1/M(b – a). Then (6.5) has a unique
solution in I and this solution is given by the absolutely and uniformly
convergent series*
b b b
f ( x )  l  K ( x, t ) f (t )dt  l 2  K ( x, t )   K (t , t1 )  f (t )dt1 dt  ... (6.8)
y( x ) 
a a a
* A convergent infinite series has a sum. But, if we differentiate (or integrate) its all the
terms, then this sum may not be equal to the derivative (or integral) of sum. However,
if the series is absolutely and uniformly convergent, then it is possible. To check this
characteristics, we check if the modulus of the nth term is less than a definite quantity.
Figure. 6.1 Region R for a  n  b and a  t  b.
Proof: Rewriting Eq. (6.7) as

b
f ( x )  l  K ( x, t1 )  y(t1 )dt1 (6.9)
y( x ) 
a
Now, replacing x by t in Eq. (6.9), we have
b
f (t )  l  K (t , t1 )  y(t1 )dt1 (6.10)
y(t ) 
a
Substituting this y(t) in R.H.S. of Eq. (6.7), we get

b
 b
y( x )  f ( x )  l  K ( x, t )  f (t )  l  K (t , t1 )  y(t1 )dt1 dt
a a 
b b b
y(x) f ( x )  l  K ( x, t )  f (t )dt  l 2  K ( x, t )  K (t , t1 )  y(t1 ) dt1dt (6.11)

a a a
Now, replacing t1 by t2 and t by t1 in Eq. (6.10), we have

b
f (t1 )  l  K (t1 , t2 )  y(t2 ) dt2 (6.12)
y(t1 ) 
a
Substituting this value of y(t1) in R.H.S. of Eq. (6.11), we get

b
y( x ) f ( x )  l  K ( x, t )  f (t ) dt 
a

b b
 b
l 2  K ( x, t )  K (t , t1 ) f (t1 )  l  K (t1 , t2 )  y(t2 ) dt2 dt1dt
a a a 
b b b
or y( x ) f ( x )  l  K ( x, t )  f (t )dt  l 2  K ( x, t )  K (t , t1 )  f (t1 ) dt1dt
a a a
b b b
 l 3  K ( x, t )  K (t , t1 )  K (t1 , t2 )  y(t2 )  dt2  dt1  dt (6.13)
a a a
Proceeding likewise, we have

b b b
y( x ) f ( x )  l  K ( x, t )  f (t )dt  l 2 a K ( x, t )a K (t, t1 ) f (t1 )dt1dt
a
b b b
   l n  K ( x, t )  K (t , t1 )  K (tn  2 , tn 1 )  f (tn 1 ) dtn 1  dt1dt
a a a
 Rn 1 ( x ) (6.14)
b b b
where, Rn+1(x) l n 1  K ( x, t )  K (t , t1 )  K (tn 1 , tn )  y(tn )dtn  dt1dt (6.15)
a a a
We now consider the following infinite series:
b b b
f ( x )  l  K ( x, t ) f (t )dt  l 2 a K ( x, t )a K (t, t1 ) f (t1 )  dt1dt   (6.16)
a
Following the assumptions (1) and (2), each term of Eq. (6.16) is continuous
in I; and it is thus obvious that Eq. (6.16) is also continuous, provided it
converges uniformly in I.
Let Un(x) represent the general term of Eq. (6.16), which we express as
b b b
U n ( x ) l n  K ( x, t )  K (t , t1 )  K (tn  2 , tn 1 )  f (tn 1 ). dtn 1  dt1dt (6.17)
a a a
Taking its modulus,

b b b
| U n ( x ) || l n  K ( x, t )  K (t , t1 )  K (tn  2 , tn 1 ) f (tn 1 ) dtn 1  dt1dt |
a a a
Therefore,
| U n ( x ) |  | l |n N  M n (b  a) n [By the conditions of theorem] (6.18)
And Eq. (6.15) converges only when
| l |  M  ( b  a)  1
or | l |  1 / M (b  a) (6.19)
and it holds good due of condition 3.
It thus means that if Eq. (6.7) has a continuous solution, it is given by
Eq. (6.14). If y(x) is continuous in I, |y(x)| must have a maximum value, say
Y. Thus,
|y(x)|  Y (6.20)
Now, from Eq. (6.15),
b b b
| Rn 1 ( x ) |  | l n 1  K ( x, t )  K (t , t1 )  K (tn 1 , tn ) y(tn ) dtn  dt1dt |
a a a
or | Rn 1 ( x ) |  | l |n 1 Y  M n 1 (b  a)n 1
Using condition 1 and Eq. (6.20) and now due to condition 3
lim Rn 1 ( x )  0
n
It thus means that function y(x) satisfying Eq. (6.14) is the continuous
function given by Eq. (6.16) or Eq. (6.8).
6.5 SOLUTION OF VOLTERRA INTEGRAL EQUATION OF THE

SECOND KIND BY SUCCESSIVE SUBSTITUTIONS
x
f ( x ) l  K ( x, t )  y(t )dt (6.21)
Theorem: Let y( x )  a
be the Volterra integral equation of the second kind. Let
1. Kernal K(x, t) / 0 be real and continuous in rectangle R for which

a  x  b, a  t  b. Also, assume
|K(x, t)|  M in R (6.22)
2. f(x) / 0, be real and continuous in the interval I, for which
a  x  b. Also, assume |f(x)|  N in I (6.23)
3. l be constant. Then Eq. (6.21) has a unique continuous solution is I and
this solution is given by absolutely and uniformly convergent series
x x t
f ( x )  l  K ( x, t ) f (t )dt  l 2  K ( x, t )   K (t1 , t2 )  f (t1 )dt1dt   (6.24)
y( x ) 
a a a
Proof: Equation (6.21) can be expressed as (let t = t1)

x
f ( x )  l  K ( x, t1 )  y(t1 ) dt1 (6.25)
y( x ) 
a
Now, replacing x by t in Eq. (6.25), we get

t
f (t )  l  K (t , t1 )  y(t1 ) dt1 (6.26)
y(t ) 
a
Substituting this value of y(t) in R.H.S. of Eq. (6.21), we obtain

x
 t
f ( x ) l  K ( x, t ) f (t )  l  K (t , t1 )  y(t1 ) dt1 dt
y( x ) 
a a 
x x t
f ( x ) l  K ( x, t )  f (t )dt  l 2  K ( x, t )
or y( x )  a K (t, t1 )  y(t1 )dt1 dt (6.27)
a a
Now, we find y(t1) to put in Eq. (6.27), and for this, we replace t1 by t2, and then,
t by t1 in Eq. (6.26), we get
t1
f (t1 )  l  K (t1 , t2 )  y(t2 )dt2
y(t1 ) 
a
Substituting this y(t1) in Eq. (6.27), we have

x
( x ) f ( x )  l  K ( x, t ) f (t )dt
y
a
 l 2  K ( x, t )  K (t , t1 )  f (t1 )  l  K (t1 , t2 )  y(t2 )dt2  dt1  dt

x t t1
a a  a 
x x t
f ( x ) l  K ( x, t )  f (t )dt  l 2  K ( x, t )  K (t , t1 )  f (t1 ) dt1  dt
or y( x ) 
a a a
x t t1
 l 3  K ( x, t )  K (t , t1 )  K (t1 , t2 )  y(t2 )dt2 dt1dt (6.28)
a a a
Proceeding similarly, we find

x x t
y( x ) f ( x )  l  K ( x, t )  f (t )dt  l 2  K ( x, t )  K (t , t1 )  f (t1 ) dt1  dt
a a a
x t tn 2
   l n  K ( x, t )  K (t , t1 )  K (tn  2 , tn 1 )  f (tn 1 ) dtn 1  dt1dt  Rn 1 ( x )
a a a
(6.29)
where,
x t tn 1
Rn 1 ( x ) l n 1  K ( x, t )  K (t , t1 )  K (tn 1 , tn )  y(tn ) dtn  dt1dt (6.30)
a a a
We now consider the following infinite series:

x x t
y( x )  l  K ( x, t )  f (t )dt  l 2  K ( x, t )   K (t , t1 ) f (t1 ) dt1   (6.31)
a a a
Now, due to conditions (1) and (2), each term of Eq. (6.31) is continuous
in I. It makes Eq. (6.31) also continuous in I, provided it converges uniformly
in I. Let Vn(x) be the general term of Eq. (6.31), given by
x t tn 2
Vn ( x )  l n  K ( x, t )  K (t , t1 )  K (tn  2 , tn 1 ). f (tn 1 ) dtn 1  dt1dt (6.32)
a a a
n
Then, | Vn ( x ) |  | l |n  N  M n ( x  a)
n!
[Here, we have applied conditions 1 and 2 over mod of Eq. (6.32)]
(b  a) n
or | Vn ( x ) |  | l |n  N  M n , a xb
n!
or | Vn ( x ) |  | l |n  N [ M (b  a)]n /n! (6.33)
Now, from Eq. (6.32), it is clear that Eq. (6.31) is convergent for all l, N,
M, (b – a); thus, from Eq. (6.33), it is followed that Eq. (6.31) is convergent
obsolutely and uniformly.
So, if Eq. (6.21) has a continuous solution, it must be expressed by
Eq. (6.29). If y(x) is continuous in I, |y(x)| must have a maximum value, say Y.
or |y(x)|  Y (6.34)
( x  a) n 1
Now, from Eq. (6.30), | Rn 1 ( x ) |  | l |n 1 Y  M n 1
(n  1)!
(b  a) n 1
| Rn 1 ( x ) |  | l |n 1 Y  M n 1 axb
(n  1)!
Hence, lim
n Rn 1 ( x ) 0
It thus follows that function y(x) satisfying Eq. (6.29) is the continuous
function given by Eq. (6.24) or Eq. (6.31). It thus proves the theorem.
6.6 SOLUTION OF FREDHOLM INTEGRAL EQUATION OF

THE SECOND KIND BY SUCCESSIVE APPROXIMATIONS:
ITERATIVE METHOD (ITERATIVE SCHEME) NEUMANN SERIES
We consider the following Fredholm integral equation of the second kind:
b
f ( x)  l 
y( x )  K ( x, t )  y(t )dt (6.35)
a
For the required solution y(x), we begin with zero order approximation.
So, let
y0(x) = f(x) (6.36)
Now, if yn(x) and yn–1(x) represent the nth and (n – 1)th order approximations
respectively, then these are connected by
b
f ( x)  l 
yn ( x )  K ( x, t )  yn 1 (t ) dt (6.37)
a
Further, the iterated kernels are given by

K1 ( x, t )  K ( x, t ) [6.38(a)]
b
and  K ( x, z)  K n 1 (z, t )dz,
K n ( x, t ) 
a
n
2,3, 4... ....[6.38(b)]
Let n = 1 in Eq. (6.37). The first order appoximation is given by

b
( x ) f ( x )  l  K ( x, t )y0 (t )dt (6.39)
y1
a
Now, since y0(t) = f(t) by [using Eq. (6.36), we have [by Eq. (6.39)]
b
( x ) f ( x )  l  K ( x, t ) f (t )dt (6.40)
y1
a
Let n = 2 in Eq. (6.37). The second order approximation y2(x) is

b
f ( x ) l  K ( x, t ) y1 (t )dt
y2 ( x ) 
a
b
f ( x ) l  K ( x, z ) y1 ( z )dz
y2 ( x )  (6.41)
a
Now, substituting y1(z) from Eq. (6.40) by replacing x by z, we get
f ( x )  l  K ( x, z )  f ( z )  l  K ( z, t ). f (t )dt  dz
b b
y2 ( x ) 
a  a 
b b b
f ( x ) l  K ( x, z )  f ( z )dz  l 2  K ( x, z )  K ( z, t )  f (t )dt dz (6.42)
y2 ( x ) 
or a a a
(also replacing z by t in the second term)

Now, interchanging the order of integration (in the last term), we have
y2 ( x ) f ( x )  l  K ( x, t )  f (t ) dt  l 2  f (t )    K ( x, z )  K ( z, t )dz  dt
b b b
a a  a 
Now, using Eqs. [6.38(a)] and [(6.38(b)] in the second and third term,
respectively, we obtain
b b
y2 ( x ) f ( x )  l  K1 ( x, t )  f (t )dt  l 2  K 2 ( x, t )  f (t )dt
a a
2 b
or f ( x )   l m  K m ( x, t )  f (t )dt (6.43)
y2 ( x ) 
a
m 1
Proceeding similarly, we generalise it and write
n b
f ( x )   l m  K m ( x, t )  f (t )dt (6.44)
yn ( x ) 
a
m 1
Now, upon taking the limit n  ¥, we find the Neumann series
 b
lim
y( x )  f ( x )   l m  K m ( x, t )  f (t )dt (6.45)
n yn ( x )  a
m 1
Resolvent Kernal (or Reciprocal Kernel)

To determine R(x, t; l) in terms of iterated kernel Km(x, t), we change the
order of integration and summation in Eq. (6.45), and obtain
b  
( x ) f ( x )  l    l m 1K m ( x, t )  f (t )dt (6.46)
y
a
 m 1 
Comparing Eq. (6.46) with
b
f ( x )  l  R( x, t; l )  f (t )dt (6.47)
y( x ) 
a

we have R( x , t; l )   l m 1K m ( x, t ) (6.48)
m 1
6.7 RESOLVENT KERNEL OF A FREDHOLM INTEGRAL EQUATION

Theorem: Let R(x, t; l) be the resolvent kernel of the following Fredholm
integral equation
b
( x ) f ( x )  l  K ( x, t ) y(t )dt
y (6.49)
a
Then
b
K ( x, t )  l  K ( x, z )  R( z, t; l )dz (6.50)
R( x , t; l ) 
a
Proof: We know (by Eq. (6.48) that


R( x , t; l )   l m 1K m ( x, t )
m 1
where the iterated kernels are

K1 ( x, t )  K ( x, t ) [6.51(a)]
b

and K m ( x, t ) 
a
K ( x, z )K m 1 ( z, t )dz, m 2, 3, ... [6.51(b)]
Then, Eq (6.48) can be expressed as


R(
x, t; l) K1 ( x, t )   l m 1K m ( x, t )
m 2
Putting values in R.H.S. using Eq. [6.51(a)] for K1(x, t) and [6.51(b)] for
Km(x, t), we get
 b
K ( x, t )   l m 1  K ( x, z )  K m 1 ( z, t ) dz
R( x , t; l ) 
a
m 2
Setting n for m – 1,
 b
K ( x, t )   l n  K ( x, z )  K n ( z, t ) dz
R( x , t; l ) 
a
n 1
 b
K ( x, t )  l  l m 1  K ( x, z )  K m ( z, t ) dz
R( x , t; l ) 
a
m 1
[Replacing n by m]
Now, interchanging the signs of summation and integration
b  
, t; l) K ( x, t )  l    l m 1K m ( z, t )  K ( x, z ) dz
R( x
a
 m 1 
Using Eq. (6.48) in the integrnal part,
b
K ( x, t )  l  [ R( z, t; l )] K ( x, z ) dz
R( x , t; l ) 
a
b
or K ( x, t )  l  K ( x, z )  R ( z, t; l ) dz
K ( x , t; l ) 
a
6.8 ILLUSTRATIONS BASED ON THE SOLUTION OF FREDHOLM

INTEGRAL EQUATION BY SUCCESSIVE APPROXIMATIONS
(ITERATIVE METHOD)
The procedure shall be in three stages. First we shall learn to find iterated
kernel, i.e., Km(x, t) from the given kernel K(x, t), refer Eq. (6.51). In the
second stage, we shall find resolvent kernel R(x, t; l), and in the third stage,
we shall solve Eq. (6.52) through R(x, t; l).
Type 1. Determination of iterated kernel

We will find the iterated kernel for
b
f ( x ) l  K ( x, t )  y (t ) dt (6.52)
y( x ) 
a
EXAMPLE 6.1: Find the iterated kernel for following kernels:

(a) K ( x,
t ) sin( x  2t ), 0  x  2p , 0  t  2p
K ( x, t ) e x cos t , 0  x  p , 0  t  p
(b) 
(c) K ( x, t )  x  sin t ,  p  x  p ,  p  t  p
(d) K ( x, t )  x  t , 0  x  1, 0  t  1
Solution: (a) Iterated kernel Kn(x, t) may be represented as [using Eq. (6.2)]
K1 ( x, t )  K ( x, t ) (i)
b
and  K ( x, z)  K n 1 z, t ) dz,
K n ( x, t ) 
a
n
2,3,... (ii)
Here, 
a 0,
b 2p .
So, K1 (
x, t ) K (
x, t ) sin( x  2t ) (iii)
Let n = 2 in Eq. (ii), we have
2p


K 2 ( x, t ) 0 K ( x, z )  K1 ( z, t ) dz
Using Eq. (iii)

2p

K 2 (
x, t ) 0 sin ( x  2 z )  sin( z  2t ) dz
1 2p
2 0

K 2 ( x, t ) [cos( x  2t  3z )  cos( x  2t  z )] dz

2p
1 1 
K 2 ( x, t )   sin( x  2t  3z )  sin( x  2t  z ) 
2  3 0

So, K 2 ( x, t )  0 (iv)
Let n = 3 in Eq. (ii). Then, we have
2p
K3 ( x, t )  0 K ( x, z )  K 2 ( z, t )dz  0 [due to Eq. (iv)]
Thus, K1(x, t) = sin(x – 2t)

and Kn(x, t) = 0, n = 2,3, ...
(b) Here, with K(x, t) = ex cos t a = 0, b = p using Eq. (i),
K 
1 ( x, t ) K ( x, t ) e x cos t (v)
and using Eq. (ii) for n = 2,
p


K 2 ( x, t ) 0 K ( x, z)  K1 (z, t )dz
for K1(z, t), we use Eq. (v).
p x


K 2 ( x, t ) 0 e cos z  e z cos t dz
p
 e x cos t .  e x  cos z dz
0
p
 e z 
 e x cos t  2 2 [cos z  sin z ]
1  1 0
1  ep x
or ( 1)1
K 2 ( x, t )  (e  cos t ) (vi)
2
Now, for K3(x, t), again using Eq. (ii) for n = 3, we get

p
K3 ( x, t )   K ( x, z ) K 2 ( z, t )dz
0
Again, using Eqs. (v) and (vi)

p  1  ep z 
0 e
x
K
3(x, t) cos z  ( 1)1  e  cos t  dz
 2 
1  ep x p

  e cos t   e z cos zdz
2 0
(1  ep ) x  1  ep 
 e cos t  ( 1)1 
2  2 

2
 1  ep  x 2
or K3(x, t) 
( 1)    e cos t (vii)
 2 
Looking the form of Eqs. (v), (vi) and (vii), we write
n 1
 1  ep 
n 1
K n ( x, t ) 
( 1)   n e x  cos t;
1, 2,3,...
 2 

(c) Here, with kernel K(x, t) = x + sin t, we have a = –p and b = p.
Now, using Eq. (i),
K1(x, t) = K(x, t) = (x + sin t) ... (viii)
For n = 2, Eq. (ii) gives
p


K 2 ( x, t ) p K ( x, z)  K1 (z, t ) dz
p
or K 2 ( x, t )  ( x  sin z )( z  sin t ) dz
p
p p p p
x  z dz  sin t  sin z dz  x sin t  dz   z sin z dz
K 2 ( x, t ) 
p p p p
p
K 2 ( x, t )  x  0  sin t  0  x sin t ( z )pp  [ z  (  cos z )]pp   1 (  cos z ) dz
p
K 2 ( x, t ) x sin t  (2p )  2p  2p (1  x sin t ) (ix)

Next, for K3(x, t), let n = 3 in Eq. (ii), using the previous results,
p p

 K ( x, z)  K2 (z, t )dz 
K3 ( x, t ) 
p  ( x  sin z){2p (1  z sin t )}dz
p
( x, t ) 4p 2 ( x  sin t )
K3 (x)
Further, for K4(x, t), let n = 4 in Eq. (ii) and using the previons results,
p


K 4 ( x, t ) p K ( x, z)  K3 (z, t )dz
p
p ( x  sin z) {4p
2
K 4 ( x, t )  ( z  sin t )}dz

K 4 ( x, t )  4p 2  2p (1  x sin t ) (xi)
or , t ) 4p 2  K 2 ( x, t ) (xii)
K 4 ( x
Similarly,  p 2 K3 ( x, t ) 16p 4 K1 ( x, t ) (xiii)
K 5 ( x, t ) 4
 p 2 K 4 ( x, t ) 16p 4 K 2 ( x, t ) (xiv)
K 6 ( x, t ) 4
In general,
If n (2p )2 m  2 ( x  sin t ), m = 1, 2, 3,...
2m  1, K 2 m 1 ( x, t ) 

K 2 m ( x, t ) (2p )2 m 1 (1  x sin t ), m = 1, 2, 3, ...
n 2m,
(d) Here, we have K(x, t) = x – t, a = 0 and b = 1.
Applying Eq. (i),
K1 ( x, t ) K ( x, t ) x  t (xv)
and by Eq. (ii) for n = 2,
1


K 2 ( x, t ) 0 K ( x, z)  K1 (z, t )dz
1
or K 2 ( x, t )  0 ( x  z) (z  t )dz [By Eq. (xv)]
x t 1
K 2 ( x, t
)   xt (xvi)
2 3
Next, for n = 3, by Eq. (ii),
1


K3 ( x, t ) 0 K ( x, z)  K2 (z, t )dz
Using Eqs. (xv) and (xvi), we have
1 z t  
K3 ( x, t
) 0 ( x  z)  2
  zt  dz
3 

After simplification,
( x  t) 1
K3 ( x, t ) 
 
 K1 ( x, t ) (xvii)
12 12
Further for n = 4, by Eq. (ii),
1


K 4 ( x, t ) 0 K ( x, z)  K3 (z, t )dz
which upon using Eqs. (xv) and (xvii) becomes
1  ( z  t ) 
K 4 ( x, t
) 0 ( x  z)   12 
 dz

Upon solving,
1  x  t 1  1
K
4 ( x, t )    xt  K ( x, t ) (xviii)
12  2 3  12 2
Proceeding similarly, we find
2
1  1 
K 5 ( x, t ) 
 K3 ( x, t ) 
  K1 ( x, t )
12 12
2
1  1 
and K 6 ( x, t ) 
 K 4 ( x, t ) 
  K 2 ( x, t )
12 12
Generalising, if n = 2m – 1, then
( 1) m 1

K 2 m 1 ( x  t ), m = 1,2,3...
12 m 1
and if n = 2m, then
( 1) m 1  x  t 1 
K2m m 1 
  xt  , m = 1,2,3...
12  2 3 
Type 2 Determination of resolvent kernel (or reciprocal kernal) R(x, t; l)

Using Eq. (6.48), the corresponding kernel for Neumann series is given by

R( x , t; l )   l m 1K m ( x, t ) (6.53)
m 1
The procedure can be followed by the example given hereunder.
EXAMPLE 6.2: Determine the resolvent kernel for the Fredholm integral
equation having kernels:
(a) K(x, t) = ex+t; a = 0, b = 1
(b) K(x, t) = (1 + x)(1 – t); a = 0, b = 1
Solution: (a) For K(x, t) = ex+t, a = 0, b = 1, [using Eq. (6.2)]
K 
1 ( x, t ) K ( x, t ) e x t (i)
1
and 
K m ( x, t ) 0 K ( x, z)  K m 1 (z, t )dz (ii)
1
Let m = 2, So, 
K 2 ( x, t ) 0 K ( x, z)  K1 (z, t )dz
1 x z
K2(x, t)  0 e  e z t dz  e x t (e2  1) / 2 (iii)
1
Let m = 3, So, 
K3 ( x, t ) 0 K ( x, z)  K2 (z, t )dz
2
1 x z  e2  1   2
x t e  1

K3(x, t) 
0 
e  e z t  2 
 
dz 
e  2  (iv)
 
An observation over Eqs. (iii) and (iv) leads to
m 1
 e2  1 
K m ( x, t )  e x t  , m = 1, 2, 3,... (v)

 2 
Thus, the required resolvent kernel R(x, t; l) given by Eq. (6.48) is,
m 1
   e2  1 

 R( x , t; l )
 l K m ( x, t )  l m 1
 m 1
.e x t
 2 
 
[Using Eq. (v)]
m 1 m 1
m 1
  l (e2  1) 
) e x  t
R( x, t; l   2  (vi)
m 1 
m 1 2
  l (e2  1)  l (e2  1)  l (e2  1) 
We find  2  
1
2

2
 +...
m 1    
l (e2  1)
As it is an infinite geometric series with the common ratio ,
2
m 1

 l (e  1) 
2
1 2
so    2



l (e  1) 2  l (e2  1)
2
(vii)
m 1 
1
2
l (e2  1) 2
Provided  1 | l | 2 (viii)
2 e 1
Finally, using Eq. (vii) in Eq. (vi), the required resolvent kernel is
 2 
l ) e x t  
R( x, t; 2  , provided Eq. (viii) is satisfied.
 2  l (e  1) 
(b) Using Eq. (6.2), for K ( x, t )  (1  x )(1  t ), a 0, b 1,
we have K1 ( x, t ) K ( x, t ) (1  x )(1  t ) (ix)
1
and 
K m ( x, t ) 0 K ( x, z)  K m 1 (z, t )dz
1
Let m = 2, So, 
K 2 ( x, t ) 0 K ( x, z)  K1 (z, t )dz
1

K 2 ( x, t ) 0 (1  x)(1  z)  (1  z)(1  t )dz
1 2
or (1  x )(1  t )  (1  z 2 ) dz  (1  x )(1  t ) (x)
0 3
Now, let m = 3, So,
1

K3 ( x, t ) 0 K ( x, z)  K2 (z, t )dz
2
1 2  2
K3(x, t)  0 (1  x)(1  z)  3 (1  z)(1  t )dz   3  (1  x)(1  t ) (xi)
Observing Eq. (x) and (xi), we express

m 1
 2
Km (
x, t )   (1  x )(1  t ) (xii)
 3
Now, the resolvent kernel is given by Eq. (6.48), which is [by Eq. (xii)]
   m 1
m 1  2 

R( x , t; l )
l
K m ( x, t )  l m 1
  
 
(1  x )(1  t )
m 1 m 1  3 
 m 1
 2l 
R( x, t; l ) (1  x )(1  t )    (xiii)
 
m 1 3
 m 1 2
 2l  2l  2l 
We find that   3  
1 
3  3 
+....
m 1
2l
is an infinite G.P.* with common ratio . Hence, this sum
3
m 1

 2l  1 3 , provided 2l  1 or | l |  3 (xiv)
    
2l 3  2l 3 2
m 1 3 1
3
Summing up, the required resolvent kernel using Eqs. (xiii) and (xiv) is
3(1  x )(1  t ) 3
R( x , t; l )  , provided | l |  (xv)
3  2l 2
Type 3: Solution of Fredholm integral equation with the help of resolvent kernel
Let
b
f ( x ) l  K ( x, t )  y(t )dt (6.54)
y( x ) 
a
be given Fredholm integral equation.
Let Km(x, t) be the mth iterated kernel and let R(x,t;l) be the resolvent
kernel for Eq. (6.54), which is given by [refer to Eq. (6.48)]

R( x , t; l )   l m 1km ( x, t ) (6.55)
m 1
Now, if the sum of the infinite series [Eq. (6.55) exists, i.e., R(x, t; l) can be
obtained in closed form (as found in Example 6.2), then the required solution
of Eq. (6.54) is given by
b
( x ) f ( x )  l  R( x, t; l ) f (t ) dt (6.56)
y
a
a
* For G.P. a + ar + ar + ...∞, S∞ = , provided | r | < 1 .
2
1− r
5x 1 1
y( x )    x.t  y(t )dt
6 2 0
by the method of successive approximation.

5x 1 1
6 2 0
y( x )
 x  t  y(t )dt (i)
Comparing Eq. (i) with Eq. (6.54), we have

5x 1
f (
x) 
,l , K ( x
, t ) xt
, a 0,
b 1
6 2
The iterated kernel Km(x, t) is given by Eq. (6.2), and we have
K 
1 ( x, t ) K ( x, t ) xt (ii)
1
and 
K m ( x, t ) 0 K ( x, z)  K m 1 (z, t )dz
1 1
Let m = 2. so, 
K 2 ( x, t ) 
0
( xz )( zt )dz
3
xt (iii)
2
1 1   1
For m = 3,  K3 ( x, t )  ( xz )  zt  dz   xt (iv)
0 3   3
m 1
 
1
Thus, we have K m ( x, t )    xt , m = 1, 2, 3.. (v)
 3
Now, by Eq. (6.48), the resolvent kernel is given by

R( x , t; l )   l m 1km ( x, t )
m 1
m 1
1  1
In our case, for l  and km ( x, t )    xt , we have
2  3
 m 1  m 1 m 1
 1  1  1
 R( x , t; l )      xt xt   
 m 1 2 3 m 1 6

 1 1   1  6 xt
R(x,t;l)  xt 1   2   xt  
 6 6   1 1 / 6 5
Finally, the required solution of Eq. (i) is given by Eq. (6.56) as
1
y( x )  f ( x )  l  R( x, t; l )  f (t )dt
0
is the present case, substituting the value, we obtain
5 x 1 1 6 xt 5t
y( x )    dt
6 2 0 5 6
5x 1 x
y( x )     x
6 2 3
EXAMPLE 6.4: By iterative method, solve

p
y( x )  1  l  sin( x  t )  y(t ) dt
0

p
y( x )  1  l  sin( x  t )  y(t ) dt (i)
0
f (
x ) 1,
l l , K ( x
, t ) sin( x  t )
th
The m iterated kernel Km(x, t) following Eq. (6.2) is given by
K1 (
x, t ) K (
x, t ) sin( x  t ) (ii)
p


K m ( x, t ) 0 K ( x, z)  K m 1 (z, t )dz
p
Let m = 2. So, 
K 2 ( x, t ) 0 K ( x, z)  K1 (z, t )dz
In the present case,
p

K2(x, t) 0 sin( x  z)  sin(z  t )dz
1 p
2 0

K 2 ( x, t ) [cos( x  t )  cos(2 z  x  t )]dz
or
p
1 1 
K
2 ( x, t ) z cos( x  t )  sin(2 z  x  t ) 
2  2 0

1 1 1  p
K 2 (
x, t )  p cos( x  t )  sin( x  t )  sin( x 
t) cos( x  t ) (iii)
2 2 2  2
Now, let m = 3, So,
p


K3 ( x, t ) 0 K ( x, z)  K2 (z, t )dz
Putting for K(x, z), and K2(x, t)
p p
K3
(x, t) 0 sin( x  z)  2 cos(z  t )dz
2
p 

which simplifies to K3 ( x, t )   sin( x  t ) (iv)
 2
3
p 
By m = 4, 
K 4 ( x, t )   cos( x  t )
 2
4
p 
and by m = 5, K 5 ( x, t )     sin ( x  t ) (v)
 2
Now, looking at Eq. (ii) to (v), the symmetry is among odd and even m, and
we express the resolvent kernel R(x,t;l) by

R( x , t; l )   l m 1K m ( x, t )
m 1
R( x, t; l ) K1 ( x, t )  l K 2 ( x, t )  l 2 K3 ( x, t ) + ...
R( x , t; l ) {K1 ( x, t )  l 2 K3 ( x, t )  l 4 K 5 ( x, t )  }
 l {K 2 ( x, t )  l 3 K 4 ( x, t )  l 5 K 6 ( x, t )  }
  lp  2  lp  4 
) sin( x  t )  1        
R( x , t; l
  2   2  
lp   lp  2  lp  4 
  cos( x  t ). 1        

2   2   2  
2
 lp  1   lp 
R( x, t;
l ) sin( x  t )  cos( x  t ) 2  , provided  2   1
 2  1  (lp / 2) 
2
 R( x , t; l ) [2 sin ( x  t )  lp cos( x  t )] (vi)
4  l 2p 2
Finally, the required solution is given by [refer to Eq. (6.56)]
p
y( x )  f ( x )  l  R( x, t; l )  f (t )dt
0
2l p
2 2 0
y( x ) 1  [2 sin( x  t )  lp cos( x – t )]dt
4l p
4l 2
y( x ) 
1 (2 cos x  lp sin x ),| l |  [After simplification]
4l p 2 2 p
EXERCISE 6.1
1/2
1. y( x ) x   y(t ) dt
0
e 1 1 1
2 2 2 0
2. y( x )  e x    y(t ) dt
1
x  l  xt  y(t ) dt
3. y( x ) 
0
x 1 p /2
4 4 0
4. y( x
) sin x   xt  y(t ) dt
1
5. y( x ) f ( x )  l  e x t  y(t ) dt
0
3 x 1 x 1 1 1
6. y( x ) e  xe    t  y(t ) dt
2 2 2 2 0
1
1  l  (1  3 xt )  y(t ) dt
7. y( x ) 
0
Answers:
1
1. x 
4
2. ex
3. (3 x ) / (3  l ),| l |  3
4. sin x.
l 1 x t
1  l 0
5. f ( x )  e f (t )dt ,| l |  1
3 x 1 x e
6. e  xe   1
2 2 3
4  2l (2  3 x )
7. ,| l |  2
4  l2
Type 4: Solution of Fredholm integral equation when the resolvent kernel

cannot be obtained in closed form
In the case when the infinite series occuring in the formulae of the resolvent
kernel cannot be determined, we use the method of successive approximations
to find the solution upto the third order.
Let the given Fredholm integral equation of the second kind be
b
f ( x ) l  K ( x, t )  y(t )dt (6.57)
y( x ) 
a
First, we take zero order approximation as
y0 ( x )  f ( x ) (6.58)
Now, if the nth order approximation is denoted by yn(x), we express
b
f ( x ) l  K ( x, t )  yn 1 (t )dt (6.59)
yn ( x ) 
a
Then, with the help of Eqs. (6.58) and (6.59), we find y1(x), y2(x) and y3(x).
Note: In case zero order approximation is provided with the problem, we
modify our results, i.e., Eqs. (6.58) and (6.59).
EXAMPLE 6.5: Solve the inhomogeneous Fredholm integral equation of

the second kind
1
y( x )  2 x  l  ( x  t )  y(t )dt
0
by the method of successive approximations to the third order by taking

y0(x) = 1.
Solution:
1
y( x )  2 x  l  ( x  t )  y(t )dt (i)
0
The zero-order approximation is

y0(x) = 1 (ii)
th
If yn(x) is n order approximation, then we apply Eq. (6.59), by which
1
2 x  l  ( x  t ) yn 1 (t )dt (iii)
yn ( x ) 
0
1
Let n = 1; y1 ( x )  2 x  l  ( x  t )  1 dt [using Eq. (ii)]
0
 1
or y1 ( x ) 2 x  l  x  
 2
Next, let n = 2. Then, by Eq. (iii),
1
y2 ( x )  2 x  l  ( x  t )  y1 (t ) dt
0
1   1
y2 ( x ) 2 x  l  ( x  t ) 2t  l  t    dt ; [By Eq. (iv)]
0
  2
 2  7
or y2 ( x )  2 x  l  x    l 2  x   , [After simplification] (v)
 3   12 
Next, let n = 3. Then, by Eq. (iii),
1
y3 ( x )  2 x  l  ( x  t )  y2 (t )dt
0
1   2  7 
y3 ( x )  2 x  l  ( x  t ) 2t  l  t    l 2  t    dt
0
  3   12  

 2  7x 2   13 x 5 
or y3 ( x )  2 x  l  x    l 2     l3   (vi)
 3  6 3  12 8 
EXERCISES 6.2
1
1. y( x ) 1  l  ( x  t )  y(t )dt
0
Answer:
 1 2 7 3  13 x 5
1. y3 (n) 
1 lx    l  x    l   
 2 12 12 8 
6.9 RECIPROCAL FUNCTIONS

The iterated kernel Kn(x, t), n = 1,2,3, ... is given by
K1 ( x, t )  K ( x, t ) [6.60(a)]
b
and 
K n ( x, t ) a K ( x, z)  K n 1 (z, t )dz [6.60(b)]
Now, let , t ) K1 ( x, t )  K 2 ( x, t )    K n ( x, t ) + ... (6.61)
 k ( x
Here, K(x, t) is real and continuous in a rectangle R, for which a  x  b
a  t  b. Let K(x, t) / 0 and let M be the maximum value of |K(x, t)| in R,
meaning by |K(x, t)|  M in R. Then, if M(b – a) < 1, it is found that the
infinite series Eq. (6.61), for k(x, t) is absolutely and uniformly convergent.
Hence, k(x,t) is real and continuous in R.
Now, (after going through Chapter 5), we know that
b

K p  q ( x, t ) a K p ( x, z)  Kq (z, t )dz (6.62)
Then, Eq. (6.61) can be expressed as
 k ( x, t )  K1 ( x , t ) K 2 ( x, t )  K3 ( x, t )    K n ( x, t )  
For K1(x, t), we use Eq. [6.60(a)] and express this as [using Eq. (6.62)]
 k ( x, t )  K ( x
, t ) K 2 ( x, t )  K3 ( x, t )    K n ( x, t )   (6.63)
b b

 a K1 ( x, z). K1 (z, t )dz  a K1 ( x, z)  K2 (z, t )dz  
b

 a K1 ( x, z)[K1 (z, t )  K2 (z, t )  ]dz
which upon using Eqs. [6.60(a)] and (6.61), becomes
b
  K ( x, z )  k ( z, t )dz (6.64)
 k ( x, t )  K ( x, t ) 
a
Again, using Eq. (6.62), Eq. (6.63) may be expressed as

b b
 k (

x, t )  K ( x, t ) a K1 ( x, z). K1 (z, t )dz  a K2 ( x, z). K1 (z, t )dz  .
b

 a [K1 ( x, z)  K2 ( x, z)  ] K1 (z, t )dz
which upon using Eqs. [6.60(a)] and (6.61), becomes
b
  k ( x.z )  K ( z, t )dz (6.65)
 k ( x, t )  K ( x, t ) 
a
Now, from Eqs. (6.64) and (6.65), we have
b
k ( x, t )  K ( x, t ) a K ( x, z)  k (z, t )dz [6.66(a)]
b
and k ( x , t )  K ( x , t ) a k ( x, z)  K (z, t )dz [6.66(b)]
Two functions K(x, t) and k(x, t) are said to be reciprocal if they both are
real and continuous in R and satisfy the conditions given in Eq. [6.66(a)] or
Eq. [6.66(b)].
Theorem: If K(x, t) is real and continuous in R, there exists a reciprocal
function k(x, t) given by
 k ( x
, t ) K1 ( x, t )  K 2 ( x, t )    K n ( x, t ) (6.67)
where K1(x, t), K2(x, t),... are iterated functions (or kernels), provided that
M(b – a) < 1, where M is the maximum value of |K(x, t)| in R for which
a  x  b a  t  b.
(Proof of this theorem has already been covered under section 6.9)
6.10 ANOTHER APPROACH TO SOLVE FREDHOLM INTEGRAL

EQUATION OF THE SECOND KIND (VOLTERRA SOLUTION)
Therem: Let the Fredholm integral equation* be
b
Now, if f ( x )   K ( x, t )  y(t )dt (6.68)
y( x ) 
a
1. K(x, t) is real and continuous in R, for which a  x  b, and a  t  b,
K(x, t) / 0;
2. f(x) is real and continuous in I (a  x  b) and f(x) / 0;
3. a function k(x, t) reciprocal to K(x, t) exists;
then the integral equation, i.e., Eq. (6.68) has a unique continuous solution
in I given by
b
y( x ) f ( x )   k ( x, t )  f (t )dt (6.69)
a
Proof: Let the variable of integration be z in place of t in Eq. (6.68).

Then, replacing x by t, we have
b
f (t )   K (t , z )  y( z )dz (6.70)
y(t ) 
a
Multiplying both sides by k(x, t) and then integrating with respect to t (a  t  b),
we have
b b b
 b
 k ( x, t ). y(t )dt a k ( x, t )  f (t )dt  a k ( x, t ) a K (t, z)  y(z)dz
a
 dt
In the last integral interchanging the order of integration, we get

b b b
 b
a k ( x, t )  y(t )dt a k ( x, t )  f (t )dt  a y(z) a k ( x, t )  K (t, z)dt  dz (6.71)
Now, since k(x, t) and K(x, t) are reciprocal functions, we have [by (6.66)b]
b
a k ( x, t ). K (t
, z )dt k ( x, z )  K ( x, z ) (6.72)
* Note that here, l is merged with K(x, t).

Using Eq. (6.72) in Eq. (6.71), we get

b b b
a k ( x, t ). y(t )dt a k ( x, t )  f (t )dt  a y(z){k ( x, z)  K ( x, z)}dz
b b b b
or, a k ( x, z)  y(z)dz a k ( x, t )  f (t )dt  a k ( x, z)  y(z)dz  a K ( x, z)  y(z)dz
b b
or 0 a k ( x, t )  f (t )dt  a K ( x, t )  y(t ) dt (6.73)
Now, from (6.69),
b
a K ( x, t )  y(t )dt y( x)  f ( x) (6.74)
Using Eq. (6.74), Eq. (6.72) can be written as
b
( x ) f ( x )   k ( x, t ) f (t )dt (6.75)
y
a
It means that if Eq. (6.68) has a continuous solution, then it is given by
Eq. (6.75), and it is unique also.
1 1 x t
2 0
EXAMPLE 6.6: Solve y( x ) f ( x)  e  y(t ) dt using Volterra method.

1 1 x t
2 0
y( x ) f ( x)  e  y(t ) dt (i)

1
K ( x, t )  e x t (ii)
2
Let k(x, t) be reciprocal kernel of K(x, t). Then, if K1(x, t), K2(x, t),... are iterated
functions, then by Eq. (6.67),
 k ( x, t )  K1 ( x, t )  K 2 ( x, t )   (iii)
And these iterated kernels are defined by Eq. (6.2), presently
1 x t
K 
1 ( x, t ) K ( x, t ) e (iv)
2
1
and 
K n ( x, t ) 0 K ( x, z)  K n 1 (z, t )dz , n = 2,3,... (v)
1
For n = 2, 
K 2 ( x, t ) 0 K ( x, z)  K1 (z, t )dz
11 1 1
0 2 e
x z
K 2 ( x, t )   e z t dz  2 e x t (vi)
2 2
1
For n = 3 
K3 ( x, t ) 0 K ( x, z)  K2 (z, t )dz
11 1 1
K3 ( x, t )  ex z  2
e z t dz  3 e x t (vii)
0 2 2 2
Looking at Eq. (v), (vi) and (vii), we can express
1
e x t , n = 1,2,3,...
K n ( x, t )  (viii)
2n
Substituting the values of K1(x, t), K2(x, t),... in Eq. (iii), we have
1 1 1  1 x t 1
t ) (e x t )   2  3  
 k ( x,   2 .e   e x t
 2 2 2  1
1
2
So, k(x, t) = – ex–t (ix)
Now, by Eq. (6.75), the solution of Eq. (i) is given by
1
y( x ) f ( x )   e x t  f (t ) dt (x)
0
and this is the required solution.
Note: The practice questions given in Exercise 6.1 can also be solved by
Volterra method.
6.11 SOLUTION OF VOLTERRA INTEGRAL EQUATION OF

THE SECOND KIND BY SUCCESSIVE APPROXIMATIONS:
ITERATIVE METHOD (NEUMANN SERIES)
Let us consider the following Volterra integral equation of the second kind:
x
f ( x ) l  K ( x, t )  y(t ) dt (6.76)
y( x ) 
a
and let the zero order approximation for the required solution y(x) be
y0(x) = f(x) (6.77)
Further, let yn(x) and yn–1(x) be nth and (n – 1)th order approximations and
these are connected by
x
f ( x ) l  K ( x, t )  yn 1 (t ) dt (6.78)
yn ( x ) 
a
We also define [refer to Eqs. [6.4(a)] and [6.4(b)] the iterated kernels as
K1 ( x, t )  K ( x, t ) [(6.79(a)]
x
and 
K n ( x, t ) t K ( x, z )  K n 1 ( z, t )dz [6.79(b)]
Now, by Eq. (6.78), for n = 1,

x
( x ) f ( x )  l  K ( x, t ) y0 (t )dt
y1
a
x
( x ) f ( x )  l  K ( x, t ) . f (t )dt [using Eq. (6.77)] (6.80)
y1
a
z
or f ( z ) l  K ( z, t )  f (t )dt [Replacing x by z] (6.81)
y1 ( z ) 
a
Again, by Eq. (6.78), for n = 2,

x
f ( x ) l  K ( x, t )  y1 (t )dt
y2 ( x ) 
a
x
or f ( x )  l  K ( x, z )  y1 ( z )dz [By replacing t by z] (6.82)
y2 ( x ) 
a
Now, substituting y1(z) from Eq. (6.81), Eq. (6.82) becomes
f ( x )  l  K ( x, z )  f ( z )  l  K ( z, t ). f (t )dt  dz
x z
y2 ( x ) 
a  a 
x
K ( x, z )  
K ( z, t ). f (t )dt  dz
x z
( x ) f ( x )  l  K ( x, z ). f ( z ) dz  l 2 
y2
or
a z a t a 
(6.83)
We now change the order of integration. Figure [6.2(a)] shows the last double
integral of Eq. (6.83). For changing the order, we refer Figure [6.2(b)] for the
same area of integration.
(a) (b)
Figure 6.2 Supporting diagrams for change of order of integration.
f ( x )  l  K ( x, z ). f ( z )dz  l 2  f (t )   K ( x, z ). K ( z, t )dz  dt
x x x
or, y2 ( x ) 
a a  t
z 
(6.84)
Now, by [6.79(b) for n = 2,
x x
K 2 ( x, t )   K ( x, z )  K1 ( z, t )dz   K ( x, z )  K ( z, t )dz [Using (6.79(a)]

t t
Now, using this result of K2(x, t) in R.H.S. of Eq. (6.84) for double integral
part, we have [replacing z by t in the second integral, along with Eq. [6.79(a)]
x x
f ( x )  l  K1 ( x, t ). f (t )dt  l 2  f (t ). K 2 ( x, t )dt
y2 ( x ) 
a a
2 x
or ( x ) f ( x )   l m  K m ( x, t ). f (t )dt (6.85)
y2
a
m 1
Proceeding similarly, we can express Eq. (6.85) as
n x
( x ) f ( x )   l m  K m ( x, t ). f (t )dt (6.86)
yn
a
m 1
Now, taking the limit n  ¥, we get the Neumann Series.

 x
y
( x) Lim
n ( x ) f ( x )   l m  K m ( x, t ). f (t )dt (6.87)
yn
a
m 1
We now determine the resolvent kernel R(x,t;l) in terms of iterated kernel
Km(x, t). Changing the order of summation and integration of Eq. (6.87), we
obtain
x  
( x ) f ( x )  l    l m 1K m ( x, t )  . f (t )dt (6.88)
y
a
 m 1 
Now, comparing Eq. (6.88) with Eq. [6.6(a)], we have

R( x , t; l )   l m 1K m ( x, t ) (6.89)
m 1
Equation (6.89) converges uniformly and absolutely when K(x, t) is continuous
in R.
6.12 RESOLVENT KERNEL AND VOLTERRA INTEGRAL EQUATION

Theorom: Let R(x,t;l) be the resolvent (or reciprocal) kernel of the Volterra
integral equation.
x
( x ) f ( x )  l  K ( x, t ). y(t ) dt
y (6.90)
a
x
then , t; l ) K ( x, t )  l  K ( x, z ). R( z, t; l )dz (6.91)
R( x
t
Proof: We know that R(x,t;l) is given by (refer to section 6.11)

R( x , t; l )   l m 1K m ( x, t ) (6.92)
m 1
where, the iterated kernels are given by
K1 ( x, t )  K ( x, t ) [6.93(a)]
x
K m ( x, t )   K ( x, z ). K m 1 ( z, t )dz
and t [6.93(b)]
Now, from Eq. (6.89), we have

R(
x, t; l ) K1 ( x, t )   l m 1K m ( x, t )
m 2
For R.H.S., using Eqs. [6.93(a)] and [6.93(b)],
 
 x
R( x
, t; l ) K ( x , t )   l m 1 t K ( x, z ). K m 1 ( z, t )dz
m 2
Now, let m –1 = n. So,
 
 x
, t; l ) K ( x , t )   l n
R( x K ( x, z ). K n ( z, t )dz
t
n 1
 x
, t; l ) K ( x, t )  l  l m 1  K ( x, z ). K m ( z, t )dz
R( x
t
m 1
Now, changing the order of summation and integration,

x
 
, t; l ) K ( x, t )  l    l m 1K m ( z, t )  K ( x, z )dz
R( x
 m 1 
t

Finally, using Eq. (6.92), we express as
x
, t; l ) K ( x, t )  l  K ( x, z ). R( z, t; l )dz (6.94)
R( x
t
6.13 ILLUSTRATIONS TO EXPLAIN THE SOLUTION OF

VOLTERRA INTEGRAL EQUATION BY SUCCESSIVE
APPROXIMATIONS (OR ITERATIVE METHOD)
Type 1: Determination of resolvent kernel of the Volterra integral equation.

x
The equation is ( x) f ( x)  l 
y K ( x, t ). y(t ) dt (6.95)
a
EXAMPLE 6.7: Find the resolvent kernel of the Volterra integral equation
with kernel K ( x, t ) 
(2  cos x ) / (2  cos t ).
Solution: We know that the iterated kernels are given by
K1 ( x, t )  K ( x, t ) (i)
x
K n ( x, t )   K ( x, z ). K n 1 ( z, t )dz , n = 2,3, (ii)
t
Here, given that K ( x, t ) = (2 + cos x ) / (2 + cos t ) (iii)

So, by Eq. (i) and (iii)
K1 ( x, t ) 
K ( x, t ) 
(2  cos x ) / (2  cos t ) (iv)
For K2(x, t), let n = 2 in Eq. (ii) , we have
x
K 2 ( x, t )   K ( x, z ). K1 ( z, t )dz
t
Now, using Eq. (iv)
x 2  cos x 2  cos z 2  cos x

K 2 ( x, t ) t .
2  cos z 2  cos t
dz
2  cos t
( x  t ) (v)
Now, let n = 3 in Eq. (ii),

x
K3 ( x, t )   K ( x, z ). K 2 ( z, t )dz
t
x 2  cos x 2  cos z
or K3 ( x, t ) t .
2  cos z 2  cos t
. ( z  t )dz [Here, we have used Eq. (iii), and (v)]
x
2  cos x  ( z  t )2  2  cos x ( x  t )2

K3 ( x, t )    (vi)
2  cos t  2  2  cos t 2!
t
Similarly, by n = 4, we find
2  cos x ( x  t )3
K 4 ( x, t )  (vii)
2  cos t 3!
After observing the values of K1(x, t), K2(x, t), K3(x, t), K4(x, t), we mention
2  cos x ( x  t ) n 1 , n = 1,2,3,... (viii)

K n ( x, t ) 
2  cos t (n  1)!
Now, by Eq. (6.92), the resolvent kernel R(x,t;l) is given by

R ( x, t ; l )   l m 1K m ( x, t ) (ix)
m 1
Expanding the R.H.S. and substituting the values of iterated kernals, we obtain
 2  cos x   l ( x  t ) l ( x  t ) 
2 2
l 3 ( x  t )3
R( x, t; l)   1     
 2  cos t   1! 2! 3! 

2  cos x l ( x t ) . This is the required resolvent kernel.
R( x , t; l )  .e
2  cos t
Type 2: Solution of Volterra integral equation with the help of resolvent kernel
Let the Volterra integral equation be
x
( x ) f ( x )  l  K ( x, t ). y(t )dt (6.96)
y
a
For its kernel K(x, t), let Km(x t) be the iterated kernels as already given in
Eqs. [6.93(a)] and [6.93(b)]. We have also learnt the method of finding the
corresponding resolvent kernel in Type 1.

R( x , t; l )   l m 1K m ( x, t ) (6.97)
m 1
Let the sum of infinite series [Eq. (6.97)] be found in closed form. Then, the
required solution of Eq. (6.96) is given by {also refer Section 6.11}
x
( x ) f ( x )  l  R( x, t; l ). f (t )dt (6.98)
y
a
EXAMPLE 6.8: By means of resolvent kernel, find the solution of

x 2  cos x
e x sin x  
y( x )   y(t )dt
0 2  cos t

x 2  cos x
y( x ) e x sin x  
 . y(t )dt (i)
0 2  cos t
2  cos x
f ( x) e x 
 sin x, l 1,
K ( x, t ) (ii)
2  cos t
Now, we refer to Example 6.7, and we find the resolvent kernel as
2  cos x x t
R( x , t; l ) 
. e (iii)
2  cos t
Then, the required solution is given by Eq. (6.98), which follows:
2  cos x x t t
x
y( x ) e x sin x  1. 
 e . e sin t dt [by Eq. (ii) and Eq. (iii)]
02  cos t
x  sin t
( x ) e x .sin x  (2  cos x )e x . 
y dt
0 2  cos t

( x ) e x sin x  (2  cos x ) e x .[log (2  cos t )]0x
y
( x ) e x sin x  (2  cos x ) e x .[log (2  cos x )  log 3]

y
  3 
y( x ) e x sin x  (2  cos x ).log 
or  
  2  cos x  
This is the required solution.
EXAMPLE 6.9: With the help of resolvent kernel, find the solution of the
following integral equation:
x 1  x2
y( x ) 1  x 2   . y(t )dt
0 1  t2
Solution:
x 1  x2
y( x ) 1  x 2   . y(t )dt (i)
0 1  t2

1  x2 , l 
f ( x)  (1  x 2 ) / (1  t 2 ) (ii)
1, K ( x, t ) 
Using Eqs. [6.93(a)] and [6.93(b)], the iterated kernels are:
K1 ( x, t )  (1  x 2 ) / (1  t 2 ) (iii)
K ( x, t ) 
x
and K m ( x, t )   K ( x, z ). K m 1 ( z, t )dz (iv)
t
Using Eqs. (ii) and (iii),
x1 x2 1  z2 1  x2

K 2 ( x, t ) t . 
1  z2 1  t2
dz
1  t2
( x  t ) (v)
x
Similarly, for m = 3, K3 ( x, t )   K ( x, z ). K 2 ( z, t )dz
t
x
x1 x2 1  z2 1  x2  ( z  t )2 
or 
K3 ( x, t ) t . 
1  z2 1  t2
( z  t ) dz
1  t2
 
 2 t
1  x 2 ( x  t )2
K3 ( x, t )  . (vi)
1  t2 2!
1  x 2 ( x  t )3

Similarly, for m 4,
K 4 ( x, t ) . (vii)
1  t2 3!
Through an observation over K1(x, t), K2(x, t), ..., we express
1  x 2 ( x  t )m 1
K m ( x, t ) 
. , m = 1,2,3,... (viii)
1  t 2 (m  1)!
Now, the resolvent kernel R(x,t;l) as given by Eq. (6.92) is

R( x , t; l )   l m 1K m ( x, t )
m 1
Expanding R.H.S. for l = 1, and substituting the values of iterated kernels,

we have
1  x 2  ( x  t ) ( x  t ) 2 ( x  t )3 
R( x, t; l ) 2 
1    
1  t  1! 2! 3! 

1  x 2 ( x t )
R( x , t; l )  e (ix)
1  t2
Finally, the required solution is given by Eq. (6.98), which in the present case is
x
( x ) f ( x )  l  R( x, t; l ). f (t )dt
y
0
x1 x2
y( x ) (1  x 2 )  1.  . e( x t ) .(1  t 2 )dt

0 1  t2
or ( x ) e x (1  x 2 ) [After simplication].
y
Type 3: Solution of Volterra integral equation when the sum of the infinite
series occuring in the formula for resolvent kernel cannot be found
Let the Volterra integral equation be
x
( x ) f ( x )  l  K ( x, t ). y(t )dt (6.99)
y
0
With the problems mentioned above, we use the following formula, known
as Neumann series:
 x
( x ) f ( x )   l m  K m ( x, t ). f (t )dt (6.100)
y
0
m 1
where, Km(x, t) is the mth iterated kernel.
EXAMPLE 6.10: Find the Neumann series for the solution of integral
equation:
x
y( x ) 1  x  l  ( x  t ). y(t )dt
0
Solution:
x
y( x ) 1  x  l  ( x  t ). y(t )dt
0 (i)
Comparing with Eq. 6.99),
f(x) = 1 + x, l = l, K(x, t) = (x – t).
Also, let Km(x, t) be the mth iterated kernel, then by Eqs. [(6.93(a)] and
[6.93(b)],
K1 ( x, t ) K ( x, t ) x  t (ii)
x
and K m ( x, t )   K ( x, z ). K m 1 ( z, t )dz (iii)
t
x
Let m = 2, So, K 2 ( x, t )   K ( x, z ). K1 ( z, t )dz
t
x
K 2 ( x, t )  t ( x  z )( z  t )dz [by Eq. (ii)]
x
 ( z  t )2  x ( 1)( z  t )2 ( x  t )3
K 2 ( x, t ) 
( x  z ).   t dz 
0 (iv)
 2  2 3!
t
Let m = 3 in Eq. (iii), we get
x
K3 ( x, t )   K ( x, z ). K 2 ( z, t )dz
t
( z  t )3 x( x  t )5
or
t  ( x  z)
K3 ( x, t ) 
3!
dz  (v)
5!
and so on. Now, the Neumann series [shown in Eq. (6.100) is
 x
( x ) f ( x )   l m  K m ( x, t ). f (t )dt
y
0
m 1
x x
y( x )  (1  x )  l  K1 ( x, t ).(1  t )dt  l 2  K 2 ( x, t ).(1  t )dt + ...
0 0
x x ( x  t )3
y( x )  (1  x )  l  ( x  t ) (1  t )dt  l 2  (1  t )dt + ...
0 0 3!
 x2 x3   x4 x5 
or y( x ) 1  x  l     l2    + ... (vi)
 2! 3!   4! 5! 
It is obtained after simplication, and this is the required solution of Eq. (i).
Remarks: 1. In this question, if l = 1, then Eq. (vi) can be summed equal to ex.
2. In this question, if we find the resolvent kernel, then

R( x, t; l )  l m 1K m ( x, t ) K1 ( x, t )  l K 2 ( x, t )  l 2 K3 ( x, t )  
m 1
( x  t )3 ( x  t )5
 l2
R( x , t; l )  ( x  t )  l 
3! 5!
Since the sum of this series cannot be found, therefore we cannot go like
Type 2.
Another example is given here in which we have adopted the procedure
of Type 3.
EXAMPLE 6.11: Solve the Volterra integral equation:

x
y( x ) 1   xt . y(t )dt
0

x
y( x ) 1   xt . y(t )dt (i)
0
Comparing Eq. (i) with Eq. (6.99), we get

f (
x ) 1,
l 1, K ( x
, t ) x.t (ii)
The mth iterated kernel, K1(x, t) = K(x, t) = xt (iii)
x
and K m ( x, t )   K ( x, z ). K m 1 ( z, t )dz (iv)
t
For K2(x, t), let m = 2 in Eq. (iv), we get

x 1 4
K 2
( x, t ) t ( xz )( 
zt )dz
3
( x t  xt 4 ) (v)
For K3(x, t) we take m = 3 in Eq. (iv), and find

x
K3 ( x, t )   K ( x, z ). K 2 ( z, t )dz
t
x 1 1 7
K3(x, t)  t
( x.z ). ( z 4 t  zt 4 ) dz 
3 18
( x t  2 x 4 t 4  xt 7 ) (vi)
1
Similarly, K 4 ( x, t
) ( x10 t  3 x 7t 4  3 x 4 t 7  xt10 ) (vii)
162
and so on.
Now, it can be verified that for the pattern of iterated kernels, the sum
of the series for the resolvent kernel cannot be ascertained; hence, we use
Neumann’s series, which is given by Eq. (6.100)
 x
( x ) f ( x )   l m  K m ( x, t ). f (t )dt (viii)
y
0
m 1
Now, expanding for m, putting the values of K1(x, t), K2(x, t), ... as found
above, and simplifying, we get
x3 x6 x9
y( x ) 
1   + ... (ix)
2 2.5 2.5.8
This is the solution of Eq. (i).
Type 4: Method of successive approximation for solving Volterra integral
equation of the second kind
Let the integral equation be
x
( x ) f ( x )  l  K ( x, t ). y(t )dt (6.101)
y
0
Also, let f(x) be continuous is [0, a] and K(x,t) be continuous for 0  x  a,
0  t  x.
We begin with some given function y0(x) continuous in [0, a]. Then,
replacing y(t) on R.H.S. of Eq. (6.101) by y0(x), we get
x
( x ) f ( x )  l  K ( x, t ). y0 (t )dt (6.102)
y1
0
Since y1(x) given by Eq. (6.102) is itself continuous in [0, a], we proceed
similarly and arrive at a sequence of functions y0(x), y1(x), y2(x), ..., yn(x), ...
x
where, ( x ) f ( x )  l  K ( x, t ). yn 1 (t ) dt
yn (6.103)
0
Because of continuity of f(x) and K(x, t), the sequence {yn(x)} converges as

n  ¥, and thus, the solution y(x) is obtained.
Remark: If we take y0(x) = f(x), we obtain Neumann series.
EXAMPLE 6.12: Using the method of successive approximation, solve the

x
1   ( x  t ). g(t )dt
g( x ) 
0
Taking g0(x) = 0.
x
1   ( x  t ). g(t )dt (i)
g( x ) 
0
Comparing Eq. (i) with Eq. (6.101) [y(x) is g(x)],
f ( x
) 1, l 1, K ( x, t
) ( x  t)
The nth order approximation,
x
( x ) f ( x )  l  K ( x, t ). gn 1 (t ). dt
gn
0
x
1   ( x  t ). gn 1 (t ). dt (ii)
gn ( x ) 
0
x
Let n = 1, 1   ( x  t ).0 dt 
g1 ( x )  1
0
Next, let n = 2 in Eq. (ii); using g1(x) = 1,
x x2 x2
g2 ( x ) 1   ( x  t ).1 dt 1  x 2 
1 
0 2 2! ,
Then, putting n = 3 in Eq. (ii) and using g2(x), we get
x  x2  x2 x4
1  ( x  t ) 1   dt 
g3 ( x )  1 
0  2!  2! 4!

In general, we have
x2 x4 x 2n 2
gn ( x ) 1    
2! 4! (2n  2)!
Note: As n  ¥, we get the infinte term to get cosh x.
Hence, the required sum is nLim
 gn ( x )  cosh x
.

following integral equation.
x
x   ( x  t ) g(t ) dt ,
g( x )  g0 (t ) 
0
0
Solution: The given equations are

x
x   ( x  t ) g(t ) dt (i)
g( x ) 
0
and g 0 (t ) = 0 (ii)
Comparing Eqs. (i), (ii) with Eq. (6.101),
[here, y( x )  g( x ) ], f ( x ) 
x, l 
1, K ( x, t ) 
x  t , g0 ( x ) 
0
The nth order approximation is given by
x
( x ) f ( x )  l  K ( x, t ). gn 1 (t ) dt
gn
0
x
or gn ( x )  x  ( 1)  ( x  t ). gn 1 (t ) dt (iii)
0
Now, for g1(x), we put n = 1 in Eq. (iii) and take g0(x) = 0.

We get g1(x) = x (iv)
For g2(x), we take n = 2, and use g1(x) = x, and obtain
x x3 x3 x3
g2 ( x ) x   ( x  t ). t dt x 
 x 
0 2 3 3!
Similarly, for g3(x), we take n = 3 in Eq. (iii) and using g2(x), we get
x3 x5
g3 ( x ) x  
3! 5!
x3 x5 x 2 n 1
In general, gn ( x )  x      ( 1) n 1
3! 5! (2 n  1)!
As the required solution is Lim
n gn ( x ) , we get
x3 x 2 n 1
Lim
n gn ( x )  x     ( 1) n     sin x
3! (2n  1)!

integral equation
x 1  g 2 (t )
g( x )   dt
0 1  t2

taking the zero order approximation (a), g0(x) = 0, and (b) g0(x) = x.
Solution: (a) The given equations is
x1 g2 (t )
g( x )   dt (i)
0 1  t2
Equation (i) is non-linear.
Comparing it with Eq. (6.101), we have
1  g2 (t )
f (
x ) 0,
l 1, K ( x, t ). g
(t )
1  t2
x
and let gn ( x )   K ( x, t ). gn 1 (t )dt
0
x1 gn21 (t )
gn ( x )   dt (ii)
0 1  t2
x 1
so that 
g1 ( x ) 
0 1  t2
dt tan 1 x
x 1  (tan 1 t )2 1

g2 ( x ) 0 1t
dt tan 1 x  (tan 1 x )3
2 3

2
 1 
1  tan 1 t  (tan 1 t )3 
x  3  dt
g3 ( x )  

0 1  t2
1 2 1
g3 ( x )  tan 1 x  (tan 1 x )3  (tan 1 x )5  (tan 1 t ) 7
3 35 79
Lim 1
Now, g( x ) n
 gn ( x ) 
tan(tan x) x
(In the expansion of tan x1, if we put x1 = tan–1 x, we get it)
(b) If we take g0 ( x ) x, g1 ( x ) x g2 ( x ) g3 ( x )  gn ( x )
Lim

So, g( x ) n
 ( x ) x
EXERCISE 6.3
1. Find the resolvent kernel, given (a) K(x, t) = 1, (b) K(x, t) = ex–t.
2. With the help of resolvent kernel, solve
x
( x ) f ( x )  l  e x t y(t )dt
(a) y
0
x
( x ) sin x  2  e x t y(t )dt
(b) y
0
x
(c) y( x ) 1   y(t )dt ,
0
x
x  (t  x ).y(t )dt
(d) y( x )  0
x
1   (t  x ).y(t )dt
(e) y( x ) 
0
x
(f) g( x ) cos x  x  2   (t  x ).g(t )dt
0

3. Using the method of successive approximation, find the solution of
x
1   y(t ) dt , y0 ( x ) 
(a) y( x ) 
 0 [RU 94, 03]
0
x
(b) y( x ) 1  x   y(t ) dt , y0 ( x ) 1
0
x
x.2   2 x t . g(t )dt , g0 ( x ) 
(c) y( x )  x
x.2 x
0
Recollect the following expansions:
2 3  n n
(a) e x 1  x  x  x     x  nLim  x
  1 
2! 3! n 0 n !  n 
x2 x3 x 4 xn
(b) log (1  x )  x       ( 1) n 1 
2 3 4 n
n 
x3 x5  np  .x ( 1) k 2 k 1
(c) sin x  x      sin    ,   x
3! 5!  2  n! k  0 (2 k  1)!
n 
x2 x4 x6  np  x x 2n
(d) cos x 
1      cos      ( 1)n
2! 4! 6!  2  n! n 0 (2n)!
3 5
(e) tan x x  x  2 x   for | x |  p
3 15 2
x3 x5 
x 2 n 1
(f) sinh x x      for all x
3! 5! n  0 (2 n  1)!

x2 x4 x 2n
(g) cosh x 1      for all x
2! 4! n  0 (2 n)!
x 3 2 x 5 17 x 7 p
(h) tanh x x      for | x | 
3! 15 315 2
1 x 3 1.3 x 5 1.3.5 x 7
(i) sin 1 x  x  .   .   ( 1  x  1)
2 3 2.4 5 2.4.6 7
3 5
(j) tan 1 x  x  x  x   ( 1  x  1)
3 5
Answers:
1. (a) exp.{l ( x  t )} , (b) exp.{( x  t )(1  l )}

x
( x ) f ( x )  l  e( x t )(1 l ) f (t )dt
2. (a) y 0
1 3x 2 1
(b) y( x )  e  sin x  cos x

5 5 5
x

(c) y = e

(d) y(x) = sin x

(e) y(x) = cos x
1
(f) g( x ) 
 cos x  sin x  x sin x
2
3. (a) y(x) = ex

(b) y(x) = 1
(c) g(x) = 2x(1–e–x)


Chapter 7
Classical Fredholm Theory
7.1 INTRODUCTION
The solution of the Fredholm integral equation of the second kind, i.e.,
b
( x ) f ( x )  l  K ( x, t ) g(t ) dt (7.1)
g
a
has been discussed in Chapter 6 as a uniformly convergent power series in

parameter l for suitably small value of | l |. As a matter of fact, Fredholm
obtained the solution of Eq. (7.1) in general form, which is valid for all values
of parameter l. These solutions are contained in three theorems, which are
known as Fredholm’s first, second and third fundamental theorems.
In this chapter, we shall study Eq. (7.1) when the functions f(x) and kernel
K(x, t) are any integrable functions. Moreover, the present method enables us
to get explicit formulae for the solution in terms of certain determinants.
7.2 FREDHOLM’S FIRST THEOREM

Statement: The non-homogeneous Fredholm integral equation of the second
kind
b
( x ) f ( x )  l  K ( x, t ) g(t ) dt (7.2)
g
a
where the functions f(x) and K(x, t) are integrable, has a unique solution
b
( x ) f ( x )  l  R( x, t; l ) f (t ) dt (7.3)
g
a
where the resolvent kernel R(x, t; l) is a meromorphic function of parameter
l defined by
D( x , t; l )
R( x , t; l )  ; [D(l) ≠ 0] (7.4)
D( l )
121
D(x, t; l) and D(l) are entire* functions of parameter l defined by Fredholm’s

series of the form

( l )m  z, z … , z m 
, t; l ) K ( x , t )  
D( x   K  1 dz1 dzm (7.5)
m 1 m !  t , z1 …, zm 

( l )m  z , ..., zm 
and D(l ) 1     K  1 dz1 dzm (7.6)
m 1 m !  z1 , ..., zm 
both of which converge for all values of l. In particular, the solution of

homogeneous integral equation is identically zero.
Moreover, it is to be noted that
K ( x1 , t1 ) K ( x1 , t2 ) K ( x1 , tn )
K ( x2 , t1 ) K ( x2 , t2 ) K ( x2 , t n )  x , x , … , xn 
K  1 2 (7.7)
: : : :  t1 , t2 , … , tn 
K ( xn , t1 ) K ( x n , t2 ) K ( xn , tn )
is known as the Fredholm determinant.

Proof: Let the interval (a, b) be partitioned into n equal parts by the points
x1 t1 a, x2 t2 a  h,..., xn tn a  (n  1) h (7.8)
where,
(b  a)
h
n
In this way, the approximation formula is developed as
b n
 K ( x, t ) g(t ) dt h  K ( x, x j ) g( x j ) (7.9)
j 1
a
Hence, Eq. (7.2) reduces to
n
( x ) f ( x )  l h  K ( x, x j ) g( x j )
g (7.10)
j 1
Since Eq. (7.10) holds for all values of x in the internal (a, b), in particular,
it must be satisfied at the n points of division xi (i = 1, 2, ..., n), and thus, the
following system of equations is obtained:
n
( xi ) f ( xi )  l h  K ( xi , x j ) g( x j ), (i = 1, 2, ..., n) (7.11)
g
j 1
Now, we express 
g( xi ) g
i , f ( xi ) fi , K
( xi , x j ) Kij (7.12)
* Refer to author’s Complex Variable by Ramesh Book Depot.

Classical Fredholm Theory 123
Then, Eq. (7.11) gives an approximation for Eq. (7.2) in terms of the system
of n linear equations with n unknowns g1, g2,..., gn as
n
gi  l h  Kij gi 
fi (i = 1, 2, ..., n) (7.13)
j 1
And Eq. (7.13) can be rewritten as
(1  l hK11 ) g1  l hK12 g2   l hK1n gn  f1 

 l hK 21 g1  (1  l hK 22 ) g2   l hK 2 n gn f2 

 (7.14)


 l hK n1 g1  l hK n 2 g2  (1  l hK nn ) gn f n 
The values of g1, g2, ..., gn obtained by solving the algebraic system Eq. (7.14)
are approximate solutions of Eq. (7.2) at the points x1, x2, ..., xn. These solutions
g1, g2, ..., gn can be plotted as ordinates, and by interpolation, we can draw a
curve g(x) which is an approximation to the actual solution.
The solutions g1, g2, ..., gn obtained by solving the algebraic system of
equations [i.e., Eq. (7.14)] may be expressed in the form of the ratios of certain
determinants, with the resolvent determinant Dn(l) of the above algebraic
system [i.e., Eq. (7.14)], where
1  l hK11  l hK12  l hK1n
 l hK 21 1  l hK 22  l hK 2 n
Dn (l )  (7.15)

 l hK n1  l hK n 2 1  l hK nn
provided that Dn(l) ≠ 0.

The approximate eigenvalues are obtained by setting this determinant to zero.
Now, expanding Dn(l) in powers of the quantity (–lh), it is found that the
first term not containing this factor is equal to unity. The term containing
(–lh) in the first power is the sum of all the determinants containing only
one column –lhKrs, r = 1, ..., n. Considering the contribution from all these
n
columns s = 1, ..., n, we find that the total contribution is  l h K ss .
s 1
The term containing the factor (–lh)2 is the sum of all the determinants
having two columns containing that factor. This gives rise to the determinants
of the form
K pp K pq
(  l h)2
K qp K qq

where (p, q) is an arbitrary pair of integers chosen from the sequence 1, ..., n
with p < q.
Similarly, the term containing the factor (–lh)3 is the sum of the
determinants of the form
K pp K pq K pr
(  l h)3 K qp K qq K qr
Krp Krq Krr

where (p, q, r) is an arbitrary triplet of integers taken from the sequence
1, 2, ..., n with p < q < r.
Proceeding likewise, we obtain the remaining terms in the expansion of
Dn(l). Thus, the determinant [Eq. (7.15)] may be expressed in the following
form:
n
(  l h)2 n K pp K pq
1  l h  K ss 
Dn (l )  
s 1  2! p, q 1 K qp K qq

K pp K pq K pr
(  l h )3 n

3!
 K qp K qq K qr 
p, q , r 1
Krp Krq Krr

K p1 p1 K p1 p2 K p1 pn
K p2 p1 K p2 p2 K p2 pn
(  l h) n n

n!

p1 , p2 ,..., pn 1

K pn p1 K pn p2 K pn pn

n
(  l h)2 n  xp xq 
1  lh  K(x p , x p ) 
Dn (l )   K 
2 ! p, q 1  x p xq 
m 1
(  l h )3 n  x p , x q , xr 
  K   (7.16)
3! p, q , r 1  x p , xq , xr 
[By applying Eq. (7.7)]

(b  a)

Since lim h lim 0 and each term of the sum shown in Eq. (7.16)
n n n
tends to be single, double, triple integral, etc., therefore we have
b bb
l2  x , x2 
1  l  K ( x, x ) dx 
D( l )    K 1 dx1 dx2
2 ! a a  x1 , x2 
a
3 b b b
l  x1 , x2 , x3 

3!    K  x1 , dx dx dx  (7.17)
x2 , x3  1 2 3
a a a
where D(l) is called Fredholm’s determinant and the series occurring on the
R.H.S. of Eq. (7.17) is called Fredhlom’s first series.
It may be pointed out that Hilbert gave a rigorous proof of the fact that
sequence Dn(l)  D(l) in the limit and Fredholm proved the convergence of
Eq. (7.17) for all values of l by using the fact that kernel K(x,t) is bounded
and is an integrable function. Thus, D(l) is an entire function of the complex
parameter l.
Now, we can find the solution of Eq. (7.2) in the form given by Eq. (7.3),
where the resolvent kernel is the quotient of D(x, t; l) and D(l). For this, we
determine D(x, t; l) as the sum of certain functional series. It is known that
the resolvent kernel R(x, t; l) satisfies the following relation:
b
, t; l ) K ( x, t )  l  K ( x, z ) R ( z, t; l ) dz (7.18)
R( x
a
From Eqs. (7.6) and (7.18), it follows that
b
D ( x , t; l ) D ( z, t; l )
 K ( x, t )  l  K ( x, z ) dz, {D(l) ≠ 0}
D (l ) a
D (l )
b
or D ( x, t; l ) K ( x, t ) D(l )  l  K ( x, z ) D( z, t; l ) dz (7.19)

a
Now, the form of the series shown in Eq. (7.17) for D(l) suggests that we
search the solution of Eq. (7.19) in the form of a power series in parameter
l, i.e.,

( l )m
x, t; l ) B0 ( x, t )  
D ( Bm ( x, t ) (7.20)
m 1 m !
For this, rewriting Eq. (7.17) as

( l )m
D (l )  1   µm (7.21)
m 1 m !
b b
 x , x ,…, xm 
where, mm    K  1 2 dx1 dxm (7.22)
a a
 x1 , x2 ,…, xm 
Now, substituting the series for D(x, t; l) and D(l) from Eqs. (7.20) and (7.21)
in Eq. (7.19) and comparing the coefficients of equal powers of l, we derive
the following recursion relations:
B0 ( x, t )  K ( x, t ) (7.23)
b
and Bm ( x, t ) µm K ( x, t )  m  K ( x, z ) Bm 1 ( z, t ) dz
 (7.24)
a
Now, we shall show that for each m(m = 1, 2, 3, ....),
b b
 x, z , z ,…, zm 
Bm ( x, t )    K  1 2 dz1 … dzm (7.25)
a a
 t , z1 , z2 ,…, zm 
First, we see that for m = 1, Eq. (7.24) takes the following form:
b
B1 ( x, t ) µ1 K ( x, t )   K ( x, z ) B0 ( z, t ) dz

a
b b
or B1 ( x, t ) K ( x, t )  K ( z, z ) dz   K ( x, z ) K ( z, t ) dz
a a
b
 x, z 
or B1 ( x, t )   K   dz (7.26)
a
 t, z 
It shows that Eq. (7.25) holds for m = 1.
To prove that Eq. (7.25) holds for general m, we expand the determinant
under the integral sign by the relation
K ( x, t ) K ( x, z1 ) K ( x, zm )
K ( z1 , t ) K ( z1 , z1 ) K ( z1 , zm )
 x , z , z ,..., zm 
K 1 1 2 
 t , z1 , z2 ,..., zm  (7.27)

K ( zm , t ) K ( zm , z1 ) K ( zm , zm )
with respect to the elements of the given row, transposing in turn the first
column one place to the right, integrating both sides and using Eq. (7.22), the
result, i.e., Eq. (7.25) follows by mathematical induction.
Now, from Eqs. (7.21), (7.23) and (7.25), we derive the so–called
Fredholm’s second series.
 b b
( l )m  x, z ,..., zm 
, t; l ) K ( x , t )  
D( x  ... K  1  dz1 ... dzm (7.28)
m 1 m ! a a  t , z1 ,..., z m 
Equation (7.28) converges for all values of l.

In the end, it will be shown that the solution in the form obtained by
Fredholm is unique and is given by Eq. (7.3). Before doing this, we find that
the integral equation [i.e., Eq. (7.18)] satisfied by R(x, t; l) is valid for all
values of l for which D(l) ≠ 0. Through Chapter 6, we already know that
Eq. (7.18) holds for (l) < B–1, where
1/2
b b 
B     | K ( x, t ) |2 dx dt 
 a a 

Since both sides of Eq. (7.18) are thus found to be meromorphic, the result
follows.
To establish the uniqueness of the solution of Eq. (7.2), it is assumed that
g(x) is a solution of Eq. (7.2), provided that D(l) ≠ 0. Rewriting (7.2) as
b
( z ) f ( z )  l  K ( z, t ) g (t ) dt (7.29)
g
a
Multiplying both sides of Eq. (7.29) by R(x, z; l), and then, integrating both
sides with respect to ‘z’ from a to b, we get
b b b b 

R ( x, z; l ) g ( z ) dz  R ( x , z ; l ) f ( z ) dz  l  
 R ( x , z ; l ) K ( z , t ) dz  g(t ) dt (7.30)

a a a a
Using Eq. (7.18), we have

b
, t; l ) K ( x, t )  l  R ( x, z; l ) K ( z, t ) dz
R ( x
a
b
or l  R ( x, z; l ) K
( z, t ) dz R ( x, t; l )  K ( x, t ) (7.31)
a
From Eqs. (7.30) and (7.31), we have

b b b
 R ( x, z; l ) g
( z )dz  R ( x, z; l ) f (z) dz   [ R( x, t; l )  K ( x, t )] g(t ) dt

a a a
b b b b
or  R ( x, t; l ) g(t )dt   R ( x, t; l ) f (t ) dt   R( x, t; l ) g(t ) dt   K ( x, t ) g(t ) dt
a a a a
b b
or  K ( x, t ) g(t ) dt   R ( x, t; l ) f (t ) dt (7.32)
a a
From Eq. (7.2), we have

b
g( x )  f ( x )
 K ( x, t ) g(t ) dt  l
(7.33)
a
and from Eqs. (7.32) and (7.33), we have

b
g( x )  f ( x )
  R ( x, t; l ) f (t ) dt
l a
b
or ( x ) f ( x )  l  R ( x, t; l ) f (t ) dt
g (7.34)
a
and this form is unique.
7.3 WORKING RULE FOR EVALUATING THE RESOLVENT KERNEL

AND SOLUTION OF FREDHOLM INTEGRAL EQUATION OF
THE SECOND KIND BY USING FREDHOLM’S FIRST THEOREM
Let the Fredholm integral equation be
b
( x ) f ( x )  l  K ( x, t ) g (t ) dt (7.35)
g
a
The Fredholm’s first theorem enables us to get explicit formulae for the
solution of Eq. (7.35) in terms of determinants.
We know that the unique and continuous solution of Eq. (7.35) is
b
( x ) f ( x )  l  R ( x, t; l ) f (t ) dt (7.36)
g
a
where the function R(x, t; l) is called Fredholm resolvent kernel of Eq. (7.35)
and is defined by
D ( x , t; l )
(7.37)
provided that D(l) ≠ 0. D (l )

( l )m
Here, , t; l ) K ( x , t )  
D ( x Bm ( x, t ) (7.38)
m 1 m !

( l )m
and D (l )  1   µm (7.39)
m 1 m !
wherein the coefficients are given by

K ( x, t ) K ( x, z1 ) K ( x, zn )
K ( z1 , t ) K ( z1 , z1 ) K ( z1 , zn )
b b

Bn ( x, t )   ... dz1 ... dzn (7.40)
a a

K ( zn , t ) K ( zn , z1 ) K ( zn , zn )
and B0(x, t) = K(x, t) (7.41)
K ( z1 , z1 ) K ( z1 , z2 ) K ( z1 , zn )
K ( z2 , z1 ) K ( z2 , z2 ) K ( z2 , z n )
b b

and µn   ... dz1 ... dzn (7.42)
a a

K ( zn , z1 ) K ( z n , z2 ) K ( zn , zn )
The function D(x, t; l) is called Fredholm minor and D(l) is the Fredholm
determinant.
In case, kernel K(x, t) is bounded or the integral
b b
 K
2
( x, t ) dx dt
a a
has a finite value, Eqs. (7.38) and (7.39) converge for all values of l, and
therefore, are entire functions for l.
Also, the resolvent kernel R(x, t; l) is an analytic function of l, except for
those values of l which are the zeros of the function D(l). Obviously, the
latter are poles of the resolvent kernel R(x, t; l).
Alternative method for calculating Bm(x,t) and mn
The coefficients mn and the function Bm(x, t) are also found from the following
recurrence relations:
We have
b
m0 = 1, µn   Bn 1 (s, s) ds, (7.43)
a
b
and Bn ( x, t ) µn K ( x, t )   K ( x, z ) Bn 1 ( z, t ) dz, m  1
 (7.44)
a
Since m0 = 1 and B0(x, t), is directly known, therefore we can use Eqs. (7.43)
and (7.44) to find, in succession µ1 , B1 ( x, t ); µ2 , B2 ( x, t ), and so on. Continuing
in this way, all the coefficients can be calculated. In certain cases, depending
on the explicit form of the kernel, Eqs. (7.38) and (7.39) contain only a finite
number of terms.
It is to be kept in mind that one distinct advantage of Fredholm method
is that Eq. (7.37) is uniformly convergent for all values of l unless D(l) = 0.
EXAMPLE 7.1: Find the resolvent kernel of the following kernels by using
Fredholm determinants:
(a) K(x, t) = xet; a = 0, b = 1
(b) K(x, t) = 2x – t; 0  x  1, 0  t  1,
(c) K(x, t) = sin x cos t; 0  x  2p, 0  t  2p,
Solution: (a) Here, K(x, t) = xet
Now, by Eq. (7.24), B0(x, t) = K(x, t) = xet
1
K ( x, t ) K ( x, z1 )
Also, by Eq. (7.26) B1 ( x, t )   dz1
0
K ( z1 , t ) K ( z1 , z1 )
1
xet xe z1
B1 ( x, t )   dz1 [Using Eq. (7.40)]
t z1
z
0 1 e z1 e
B1(x, t) = 0
(since the columns of the determinat under the integral sign are identical)
11 K ( x, t ) K ( x, z1 ) K ( x, z2 )
Similarly, B2 ( x, t )    K ( z1 , t ) K ( z1 , z1 ) K ( z1 , z2 ) dz1dz2
00
K ( z2 , t ) K ( z2 , z1 ) K ( z2 , z2 )

xet xe z1 xe z2
11

 z1e z1e z1e dz1dz2
t z1 z2
B2 ( x, t ) 0
00
z2 et z2 e z1 z2 e z2

Since B1(x, t) = B2(x, t) = 0, it follows that Bn(x, t) = 0, for n  1. Thus, from
Eq. (7.42), we have
1 1

K ( z1 , z1 )dz1  z1e z1
m1 dz1
0 0
1
m1 [ z1e z1 ]10   e z1 dz1
0
m1 e  [e z1 ]10 e  (e  1) 1
11
z1e z1 z1e z2
and m2   z ez1 z ez2 dz1dz2 0

00 2 2
Clearly, mm = 0 for all m  2.

l2
Now, by Eq. (7.5), D( x, t; l ), K ( x, t )  l B1 ( x, t )  B ( x, t ) 
2! 2
D(x, t; l) = xet (By substituting values of K(x, t) B1(x, t), etc.)
l2
and by Eq. (7.6), D(l ) 1  lm1  m2  1  l
2!
(By substituting the values of m1, m2, etc.)
Thus, the Fredholm resolvent kernel as given by Eq. (7.4) is
D( x , t; l ) xet

R ( x , t; l ) 
D( l ) 1 l
(b) Here, we have B0(x, t) = K(x, t) = 2x – t; and then, similar to part (a),
1 1
K ( x, t ) K ( x, z1 ) 2 x  t 2 x  z1
 B1 ( x, t )  dz
K ( z1 , t ) K ( z1 , z1 ) 1  2z1  t 2 z1  z1
dz1
0 0
1
B1 (
x, t )  [ z1 (2 x  t )  (2 x  z1 )(2 z1  t )]dz1
0
1
 [ z1 (2 x  t )  (4 xz1  2 xt  2z1  z1t )] dz1

2
or B1 (
x, t )
0
1
 [2z1  z1 (2 x  t  4 x  t )  2 xt ] dz1
2
or B1 ( x
, t)
0
1
 2z3 z 2  2
or B1 ( x, t )   1  1 ( 2 x  2t )  2 xtz1    x  t  2 xt
 3 2 0 3
11 K ( x, t ) K ( x, z1 ) K ( x, z2 )
Also, B2 ( x, t )    K ( z1 , t ) K ( z1 , z1 ) K ( z1 , z2 ) dz1dz2
00
K ( z2 , t ) K ( z2 , z1 ) K ( z2 , z2 )
11 2x  t 2 x  z1 2 x  z2
B2 ( x, t )    2z1  t 2 z1  z1 2 z1  z2 dz1dz2
00 2 z2  t 2 z2  z1 2 z2  z2

or B2(x, t) = 0. (After little simplification)
Hence, Bp(x, t) = 0 for all p  2.
1 1 1
1  1
Again, m1   K ( z1 , z1 ) dz1   (2 z1  z1 )dz1   z12  
0 0  2 0 2
11
K ( z1 , z1 ) K ( z1 , z2 )
and m2    dz dz
00
K ( z2 , z1 ) K ( z2 , z2 ) 1 2
11
2 z1  z1 2 z1  z2 1

m2 
 2z2  z1 2 z2  z2 dz1dz2 3
(Upon simplification)
00
111 K ( z1 , z1 ) K ( z1 , z2 ) K ( z1 , z3 )
and m3     K ( z2 , z1 ) K ( z2 , z2 ) K ( z2 , z3 ) dz1dz2 dz3
000 K ( z3 , z1 ) K ( z3 , z2 ) K ( z3 , z3 )
111 2 z1  z1 2 z1  z2 2 z1  z3
m3     2 z2  z1 2 z2  z2 2 z2  z3 dz1dz2 dz3
000 2 z3  z1 2 z3  z 2 2 z3  z 3

m3 = 0 (Upon simplification).
Hence, mp = 0 for all p  3.
Thus, we have upon using Eq. (7.5),

( l )m
K ( x, t )  
D( x , t; l )  Bm ( x, t ) 
K ( x, t )  l B1 ( x, t )
m 1 m !
2 
D( x, t; l ) 2 x  t  l   x  t  2 xt 
3 

and by using Eq. (7.6),

( l )m l2 1 l2
1 
D(l )  m m 
1 lm  m2 
1 l
m 1 m ! 2! 2 6
The Fredholm resolvent kernel is given by Eq. (7.4) providing
2 
2 x  t  l   x  t  2 xt 
D( x , t; l ) 3 

R ( x , t; l ) 
D( l ) 1  (l /2)  (l 2 /6)
(c) Here, B0(x, t) = K(x, t) = sin x cos t. As before
2p
K ( x, t ) K ( x, z1 )
B1 ( x, t )   K ( z1 , t )
dz
K ( z1 , z1 ) 1
0
2p
sin x cos t sin x cos z1
B1 ( x, t )   dz = 0
sin z1 cos t sin z1 cos z1 1
0

2p 2p
Next, m1 
K ( z1 , z1 )dz1  sin
z1 cos z1 dz1 0
0 0

Thus, by using Eq. (7.5), we have

( l )m
, t; l ) K ( x , t )  
D( x Bm ( x, t )
m 1 m !
D( x , t; l ) K ( x, t )  l B1 ( x, t ) 
D(x, t; l) = sin x cos t (By substituting the above values)
By using Eq. (7.6),

( l )m
D(l ) 1   mm
m 1 m !
D(l ) 1  lm1  1
(Upon substituting the above values)
The Fredholm resolvent kernel is given by Eq. (7.4), providing

D( x , t; l )

R ( x , t; l )  sin x cos t
D( l )
EXAMPLE 7.2: Find the resolvent kernel of the following kernels by using
the Fredholm determinants:
K(x, t) = 1 + 3xt, 0  x  1, 0  t  1
(a)
K(x, t) = x2t – xt2, 0  x  1, 0  t  1
(b)
Solution: (a) Here, B0(x, t) = K(x, t) = 1 + 3xt. Now,
1
K ( x, t ) K ( x, z1 )
B1 ( x, t )   dz
K ( z1 , t ) K ( z1 , z1 ) 1
0
1 1  3 xt 1  3 xz1
B1 ( x, t )   dz1
1  3z1t 1  3z12
0
 3( x  t ) 
B1 ( x, t ) 
  3 xt  1 (Upon simplification)
 2 
1 K ( x, t ) K ( x, z1 ) K ( x , z2 )
Also, B2 ( x, t )   K ( z1 , t ) K ( z1 , z1 ) K ( z1 , z2 ) dz1 dz2

0 K ( z2 , t ) K ( z2 , z1 ) K ( z2 , z2 )
1  3 xt 1  3 xz1 1  3 xz2
11
B2 ( x, t )    1  3z1t 1  3z12 1  3z1 z2 dz1 dz2
00
1  3z2 t 1  3z2 z1 1  3z22

B2(x, t) = 0 (Upon simplification)
1 1
 K (z1 , z1 )dz1 
 (1 3z1 ) dz1 
2
Next, m1  2
0 0
11
K ( z1 , z1 ) K ( z1 , z2 )
m2    dz dz
K ( z2 , z1 ) K ( z2 , z2 ) 1 2
00
11
1  3z12 1  3z1 z2
m2    dz1dz2
00 1  3z2 z1 1  3z22

1
m2  , (upon simplification)
2
Also, m3 = 0 (can be verified by readers)

( l )m
Now, we have , t; l ) K ( x , t )  
D( x Bm ( x, t )
m 1 m !
 3( x  t ) 
D( x, t; l ) 
1 3 xt  l   3 xt  1
 2 

( l )m
and D(l ) 1   mm
m 1 m !
l2 l2
D(l ) 1  lm1 
m2  1  2l 
2! 4
The Fredholm resolvent kernel is given by
D( x , t; l )
R( x , t; l ) 
D( l )
 3( x  t ) 
1  3 xt  l   3 xt  1
 2 
R( x , t; l ) 
1  2l  (1 / 4)l 2
(b) Here, B0(x, t) = K(x, t) = x2t – xt2. Then, as before
1
K ( x, t ) K ( x, z1 )
B1 ( x, t )   dz
K ( z1 , t ) K ( z1 , z1 ) 1
0
1
x 2 t  xt 2 x 2 z1  xz12
B1 ( x, t )   dz1
0 z12 t  z1t 2 z12 z1  z1 z12

 x  t xt 1 
B1 ( x, t ) 
 xt    
 4 3 5

Also, B2(x, t) = 0
1 1
Next, m1  K ( z1 , z1 )dz1  ( z13  z13 )dz1 0
0 0
11
K ( z1 , z1 ) K ( z1 , z2 )
m2    dz dz
K ( z2 , z1 ) K ( z2 , z2 ) 1 2
00
11
0 z12 z2  z1z22
m2    dz1dz2
00 z22 z1  z2 z12 z22 z2  z2 z22

1
m2  , (Upon simplification).
120
Also, we find m3 = 0.
Now, we have

( l )m
, t; l ) K ( x , t )  
D( x Bm ( x, t )
m 1 m !
D( x , t; l ) K ( x, t )  l B1 ( x, t ) 
 x  t xt 1 
D( x, t; l )  x 2 t  xt 2  l xt    
 4 3 5


( l )m l2
and 1 
D( l )  mm 
1  lm1  m2 
m 1 m ! 2!
l2
D(l ) 1 
240
 x  t xt 1 
x 2 t  xt 2  l xt    
D( x , t; l )  4 3 5

Finally, R ( x , t; l ) 
D( l ) 2
1  (l /240)
EXAMPLE 7.3: Using the recurrence relations, find the resolvent kernels of
the following kernels:
(a) K(x, t) = x – 2t; 0  x  1, 0  t  1
(b) K(x, t) = sin x cos t; 0  x  2p, 0  t  2p
(c) K(x, t) = 4xt – x2; 0  x  1, 0  t  1
Solution: (a) Here, K(x, t) = x – 2t.
The resolvent kernel R(x, t; l) is given by Eq. (7.4) as
D( x , t; l )
R( x , t; l ) 
D(l ) (i)

( l )m
, t; l ) K ( x , t )  
where by Eq. (7.5), D( x Bm ( x, t ) (ii)
m 1 m !

( l )m
and by Eq. (7.6), D(l ) 1   m m (iii)
m 1 m !
Now, we have B0(x, t) = K(x, t) = x – 2t (iv)

Also, from Eqs. (7.43) and (7.44),
1
m0 = 1 and m p   Bp 1 (s, s) ds , p  1 (v)
0
1
and Bp ( x, t ) m p K ( x, t )  p  K ( x, z ) Bp 1 ( z, t ) dz , p  1
 (vi)
0
Letting p = 1 in Eq. (v), we obtain
1 1
1
 B0 (s, s) ds 
m1   (s  2s)ds 

2
0 0
Let p = 1 in Eq. (vi). We obtain

1
B1 ( x, t ) m1K ( x, t )   K ( x, z ) B( z, t )dz

0
1
1
 ( x  2t )   ( x  2 z )( z  2t ) dz
B1 ( x, t ) 
2
0
1
1
B1 ( x, t )  ( x  2t )   {2 z 2  z ( x  4t )  2 xt} dz
2
0
1
1  2 z 3 z 2 
B1 ( x, t ) 
 ( x  2t )    ( x  4t )  2 xtz 
2  3 2 
0
1  2 1 
B1 ( x, t )  ( x  2t )     ( x  4t )  2 xt 
2  3 2 
2
B1 ( x, t ) 
 2 xt  x  t
3
Now, taking p = 2 in Eq. (v), we obtain
1 1
2 
 B1 (s, s) ds    3  2s
2
m2   2s  ds

0 0
1
2 2s3  1
m2   s   s2  
 3 3 0 3
Next, taking p = 2 in Eq. (vi), we obtain
1
B2 ( x, t ) m2 K ( x, t )  2  K ( x, z ) B1 ( z, t ) dz
and 
0
1
1 2 
B2 ( x, t
) ( x  2t )  2  ( x  2 z )   2 zt  z  t  dz
3 3 
0
1
1  2x 4 
B2 ( x, t ) ( x  2t )  2    2 xzt  xz  xt  z  4 z 2 t  2 z 2  2 zt  dz
3 0
3 3 

1
1  2 xz xz 2 2 4 2 
B2 ( x, t )  ( x  2t )  2   xtz 2   xtz  z 2  z 3t  z 3  z 2 t 
3  3 2 3 3 3 
0
1 4x 4 8 4
B2 ( x, t ) ( x  2t )   2 xt  x  2 xt   t   2t
3 3 3 3 3
B2 ( x, t )  0
Since B2(x, t) = 0, therefore by Eq. (vi),
Bp(x, t) = 0 and mp = 0 for all p  3.
Substituting the above values in Eq. (ii) and (iii), we have
D( x , t; l ) K ( x, t )  l B1 ( x, t ) 
2 
D( x, t; l )  x  2t  l   2 xt  x  t )
3 

l2
and D( l ) 
1  lm 1  m 
2! 2
l l2
D(l ) 1  
2 6
2 
x  2t  l   2 xt  x  t 
3 
Thus, by Eq. (i), R( x, t; l ) 
l l2
1 
2 6
(b) Here, K(x, t) = sin x cos t.
Now, B0(x, t) = K(x, t) = sin x cos t
2p
Also, m0 = 1, m p   Bp1 (s, s)ds, p  1 (vii)
0
2p
and Bp ( x, t ) m p K ( x, t )  p  K ( x, z ) Bp 1 ( z, t )dz , p  1
 (viii)
0
Let p = 1, in Eq. (vii). We have
2p 2p
m1 
B0 (s, s)ds  sin s cos s ds
0 0
2p 2p
1 1   cos 2s 
m1  
2 0
sin 2 s ds  0
2  2  0

Now, taking p = 1 in Eq. (viii), we obtain
2p

B1 ( x, t ) m1K ( x, t )   K ( x, z)B0 (z, t ) dz
0
2p
B1 ( x, t )    (sin x cos z )(sin z cos t )dz
0
2p
 sin x cos t  sin z cos z dz 
B1 ( x, t )  0
0
Since B1(x, t) = 0, therefore by recurrence relations,

Substituting these values in Eq. (ii) and (iii), we have
D( x
, t; l ) K
( x, t ) sin x cos t
and D(l) = 1
Hence, R(x, t; l) = sin x cos t
(c) Here, K(x, t) = 4xt – x2.
We have B0(x, t) = K(x, t) = 4xt – x2
1
m0 = 1 and m p   Bp 1 (s, s)ds , p  1 (ix)
0
1
 (x)
0
Taking p = 1 in Eq. (ix), we obtain
1 1
m1   B0 (s, s)ds   (4s 2  s 2 )ds [ s 3 ]10 1
0 0
Now, taking p = 1 in Eq. (x),we obtain

1
B1 ( x, t ) m1 K ( x, t )   K ( x, z ) B0 ( z, t )dz

0
1
B1 ( x, t )  4 xt  x 2   (4 xz  x 2 )(4 zt  z 2 ) dz
0
1
 z3 
) 4 xt  x    xz 4  ( x 2  16 xt )  2 x 2 tz 
B1 ( x, t  2
 3  0

 1 
B1 ( x, t ) 4 xt  x 2    x  ( x 2  16 xt )  2 x 2 t 
 3 
4 4
B1 ( x, t) 2 x 2 t  x 2  x  xt
3 3
Next taking, p = 2 in Eq. (ix), we obtain
1 1
 4 2 4 
 B1 (s, s   2s
3

m2 )ds  s  s  s 2  ds
3 3 
0 0
1
s 4 4
s 2 4s3  1
m2    s 3    
 2 9 2 9 0 9

and now putting p = 2 in Eq. (x), we obtain
1
B2 ( x, t ) m2 K ( x, t )  2  K ( x, z ) B1 ( z, t ) dz

0
1
1  4 4 
B2 ( x
, t) (4 xt  x 2 )  2  (4 xz  x 2 )  2 z 2 t  z 2  z  zt  dz
9  3 3 
0
1
1   4  4t  
B2 ( x, t
) (4 xt  x 2 )  2  (4 xz  x 2 )  z 2  2t    z 1    dz
9 0   3  3 
1
1   4   4t   4  4t  
B2 ( x, t ) (4 xt  x 2 )  2   4 x  2t   z 3  z 2 4 x 1    x 2  2t     x 2 1   z  dz
9   3    3   3   3 
0
1
1   4 z 3   4t   4   x 2  4t  
B2 ( x, t ) = (4 xt − x 2 ) − 2  x  2t −  z 4 + 4 x 1 −  − x 2  2t −   − 1 −  z 2 
9   3 3  3  3  2  3  0
1   4 1   4t   4   x2  4t  
B2 ( x, t
) (4 xt  x 2 )  2  x  2t    4 x 1    x 2  2t     1   
9   3  3  3   
3  2 3 
B2(x, t) = 0
Since B2(x, t) = 0, therefore Bp(x, t) = 0 and mp = 0 for all p  2.
Substituting these values in Eq. (ii) and (iii), we have
 4 4 
D( x, t; l ) K ( x, t )  l B1 ( x, t ) 4 xt  x 2  l  2 x 2 t  x 2  x  xt 
 3 3 
l2 l2
and D(l ) 1  lm1  m2 1  l 
2! 18
 4 4 
4 xt  x 2  l  2 x 2 t  x 2  x  xt 
 3 3 
and thus, R ( x, t; l ) 
1  l  (l 2 / 18)
EXAMPLE 7.4: Determine D(l) and D(x, t; l) for the following kernels for
the prescsibed limits a and b:
(a) K(x, t) = 1; a = 0, b = 1
(b) K(x, t) = sin x; a = 0, b = p
Solution: (a) Here, K(x, t) = 1.

We know that

( l )m
, t; l ) K ( x , t )  
D( x Bm ( x, t ) (i)
m 1 m !

( l )m
and D(l ) 1   m m (ii)
m 1 m !
B0(x, t) = K(x, t) = 1 (iii)

1
m0 = 0 and m p   Bp 1 (s, s) ds, p  1 (iv)
0
1
 (v)
0
Taking p = 1 in Eq. (iv), we obtain

1 1

m1  B0 (s, s)
ds 
ds 1
0 0
and taking p = 1 in Eq. (v), we obtain

1
B1 ( x, t ) m1K ( x, t )   K ( x, z ) B0 ( z, t )dz

0
1
B1 ( x, t )  1   dz  1  1  0
0
Since B1(x, t) = 0, therefore Bp(x, t) = 0 and mp = 0 for all p  2.

Substituting the above values in Eq. (i) and (ii), we obtain
D( x , t ; λ ) = K ( x , t ) = 1

D( λ ) = 1 − λ
and
(b) Here, K(x, t) = sin x, Now, following the approach of part (a), we have
B0(x, t) = K(x, t) = sin x.
p
Again, m
0 = 0 and µp  Bp 1 (s, s) ds, p  1 (vi)
0
p
and µp K ( x, t )  p  K ( x, z ) Bp 1 ( z, t ) dz, p  1 (vii)
Bp ( x, t ) 
0
Taking p = 1 in Eq. (vi), we obtain
p p
 B0 (s, s) ds 
µ1   sin s ds 
[  cos s ]p0 
2
0 0
Taking p = 1 in Eq. (vii), we obtain
p
B1 ( x, t ) µ1K ( x, t )   K ( x, z ) B0 ( z, t ) dz

0
p
( x, t ) 2 sin x   sin x sin z dz
B1
0
B1 ( x, t )  2 sin x  sin x [  cos z ]p0
B1 ( x, t ) 2 sin x  2 sin x 0
Since B1(x, t) = 0, therefore by recursive nature,

Substituting these values in Eq. (ii) and (iii), we get
D ( x , t; l ) 
K ( x, t ) 
sin x, and D (l ) 
1  l µ1 
1  2l
EXAMPLE 7.5: Using Fredholm theory, solve the following integral

equations:
10
(a) g ( x) e x  l  xt g (t ) dt
0
1
g ( x ) f ( x )  l  ( x  t ) g (t ) dt
(b)
0
10
Solution: (a) Given g ( x) e x  l  xt g (t ) dt (i)
0
Comparing Eq. (i) with
10
( x ) f ( x )  l  K ( x, t ) g (t ) dt
g
0
we get f(x) = ex and K(x, t) = xt

We know that B0(x, t) = K(x, t) = xt
10

m0 = 1 and µp  Bp 1 (s, s) ds , p  1 (ii)
0
10
and µp K ( x, t )  p  K ( x, z ) Bp 1 ( z, t ) dz , p  1 (iii)
Bp ( x, t ) 
0
Taking p = 1 in Eq. (ii), we obtain
10 10 10
 s3  103
µ1  B0 (s
, s) ds s ds   
2
0 0  3 0 3
Now, taking p = 1 in Eq. (iii), we obtain

1
B1 ( x, t ) m1K ( x, t )   K ( x, z ) B0 ( z, t ) dz

0
3 10
10
B1
( x, t ) xt   ( xz )( zt ) dz
3
0
10
103  z3 
B1 ( x, t )  xt  ( xt )    0
3  3 0

Since B1(x, t) = 0, therefore Bp(x, t) = 0, m1 = 0 for all p  2.
Substituting these values in Eqs. (7.38) and (7.39), we have
D( x
, t; l ) K
( x, t ) xt
 103 
D( l )  1  lm1  1 l 
 3 
xt
R ( x , t; l ) 
1  l (103 /3)
Thus, the required solution of Eq. (i) is given by
10
( x ) f ( x )  l  R( x, t; l ) f (t ) dt
g
0
10
lx
) ex   te dt
t
g( x 3
1  l (10 / 3)
0
lx
ex 
g( x )  [tet  et ]10
0 (Integrating by parts)
1  l (103 / 3)
3l x
g( x ) e x  (10e10  e10  1) (Upon simplification)
3  1000 l
3l x
ex 
g( x )  (1  9e10 )
3  1000 l
(b) Here, we have a = 0, b = 1 and K(x, t) = x + t
Again, B0(x, t) = K(x, t) = x + t
1
m0 = 1 and m p   Bp 1 (s, s)ds , p  1 (iv)
0
1
and Bp ( x, t ) m p K ( x, t )  p  K ( x, z ) Bp 1 ( z, t )dz (v)

0
For p = 1, Eq. (iv) gives

1 1

m1  B0 (s, 
s)ds  2
s ds s 2 ]10 1
[
0 0
Again, for p = 1, Eq. (v) provides

1

0
1
B1 ( x, t )  ( x  t )   ( x  z )( z  t ) dz
0
1
B1 ( x, t )  ( x  t )   [ z 2  z ( x  t )  xt ] dz
0
1
1 z2 
B1 ( x, t )  ( x  t )   z 3  ( x  t )  xtz 
 3 2 0
1 1
B1 ( x, t )  ( x  t ) 
 ( x  t )  xt
3 2
1 1
B1 ( x, t ) ( x  t )  xt 
2 3
Also, putting p = 2 in Eq. (iv), we obtain
1 1
1 1
 B1 (s, s) ds   2 (s  s)  s
2
m
2    ds
3
0 0
1
 s 2 s3 1  1
m2 
   s  
 2 3 3  6
0
Now, puttng p = 2 in Eq. (v), we obtain

1
B2 ( x, t ) m2 K ( x, t )  2  K ( x, z ) B1 ( z, t )dz

0
1
1 1 1
B2 ( x, t )  ( x  t )  2  ( x  z )  ( z  t )  zt   dz
6 2 3
0
1
1  1  1 1 x  1 1 
B2 ( x, t )  ( x  t )  2   z 2   t   z  t    xt   x  t    dz
6 2  2 3 2  2 3 
0
1
1  z3  1  z2 1 1 x   t 1 
B2 ( x, t )  ( x  t )  2    t    t    xt   xz    

6  3  2  2 2 3 2 2 3 
0
1 1  1  1  1 1 x   t 1 
B2 ( x, t )  ( x  t )  2    t    t    xt   x    
6 
3 2  
2 2 3 2   2 3 

B2 ( x, t ) = 0 (Upon simplification)
Since B2(x, t) = 0 , it follows from Eqs. (iv) and (v), that

Bp(x, t) = 0, m0 = 0 for all p  3
Therefore, D( x
, t; l ) K ( x, t )  l B1 ( x, t )
1 1
D( x, t; l )  x  t  l  ( x  t )  xt  
 2 3
l2 1
and D(l ) 1  lm1  m2 1  l  l 2
2! 12
1 1
x  t  l  ( x  t )  xt  
 2 3
and thus, R ( x , t; l )  2
1  l  (l / 12)
Hence, the solution of the integral equation is given by
1
( x ) f ( x )  l  R( x, t; l ) f (t )dt
g
0
1 1
( x  t )  l  ( x  t )  xt  
1
( x) f ( x)  l 
g 2 3
f (t )dt .
or 2
0 1  l  ( l / 12)
EXAMPLE 7.6: Solve the following integral equations:

1
x  l  [ xt  ( xt )1/2 ]g(t )dt
(a) g( x ) 
0
1
x
(b) g( x )   l  (2 x  t ) g(t ) dt
6 0
Solution: (a) The given integral equation is
1
x  l  [ xt  ( xt )1/2 ] g(t ) dt (i)
g( x ) 
0
Comparing it with
b
( x ) f ( x )  l  K ( x, t ) g(t ) dt
g
a
we find, a = 0, b = 1, f(x) = x and K(x, t) = xt + (xt)1/2

so that here we have
K ( x, t ) xt  ( xt )1/2 , m0 
B0 ( x, t )  1
1 1 1
 s3 s 2  5
and m1  B0 (s, s) ds  (s  s) ds    
2
0 0  3 2  6
0
1
Next, 
0
1
5
B1 ( x, t )  {xt  ( xt )1/2 }   {xz  ( xz )1/2 }{zt  ( zt )1/2 }dz
6
0
1
5  xtz 3 z 5/2 z 2 
B1 ( x, t )  {xt  ( xt )1/2 }    (x t  t x )  ( xt )1/2 
6  3 5/2 2 
0
5  xt 2 1 
B1 ( x, t )  {xt  ( xt )1/2 }    ( x t  t x )  ( xt )1/2 
6 3 5 2 
1 1 2
B1 ( x, t ) xt  ( xt )1/2  ( x t  t x )
2 3 5
1
Then, m2   B1 (s, s) ds
0
1
1 1 2 
  2 s
2
m
2  s  (s s  s s)  ds
3 5 
0
1
1 1 4 
  2 s
2
m
2  s  s 3/2  ds
3 5 
0
1
 s 3 s 2 4 s 5/2  1
m2       
 6 6 5 5 / 2  75
0
1
B2 ( x, t ) m2 K ( x, t )  2  (K ( x, z ) B1 ( z, t ) dz
Further, 
0
1
1 1 1 2 
B2 ( x, t )  {xt  ( xt )1/2 }  2  {xz  ( xz )1/2 }  zt  ( zt )1/2  ( z t  t z )  dz
75 0 2 3 5 
1  xtz 3 2 x tz 5/2 2 x tz 3 4 xtz 5/2 t xz 5/2

B2 ( x, t )  {xt  ( xt )1/2 }  2     
75  6 15 15 25 5
1
( xt )1/2 z 2 4( xt )1/2 5/2 t x z 2 
  z  
6 25 5 
0
1  xt 2 x t 2 x t 4 xt t x ( xt )1/2  4  
B2 ( x, t )  {xt  ( xt )1/2 }  2          ( xt )1/2  t x 
75  6 15 15 25 5 6  25  
B2 ( x, t ) = 0 (Upon simplification)
Since B2(x, t) = 0, therefore we infer that

Bp(x, t) = 0, mp = 0 for all p  3
Thus, we find that
D( x
, t; l ) K ( x, t )  l B1 ( x, t )
1 1 1 
xt  ( xt )1/2  l  xt  ( xt )1/2  ( x t  t x )
D( x , t; l ) 
2 3 2 
and
l2 5 1 2
D (l ) 
1 l c1  c2 
1 l l
2! 6 150
Therefore, the resolvent kernel is given by
1 1 2 
xt  ( xt )1/2  l  xt  ( xt )1/2  ( x t  t x 
 2 3 5 
R ( x , t; l ) 
1  (5/6)l  (1/150)l 2
Hence, the required solution of the given equation is
1
( x ) f ( x )  l  R( x, t; l ) f (t ) dt
g
0
1 1 2 
xt  ( xt )1/2  l  xt  ( xt )1/2  ( x t  t x )
1
2 3 5  dt
or g( x ) x  l  2
0 1  (5 / 6)l  (1 / 150)l
l
g( x ) x 
1  (5 / 6)l  (1/150)l 2
1
 xt 3 2 xt 5/2 l xt 3 2 x l t 5/2 4 x l t 5/2 2 x l t 3 
      
 3 5 6 15 25 15 
0
l  x 2 x l x 2l x 4 x l 2l x 
g( x ) x  2 3
     
1  (5/6)l  (1/150)l  5 6 15 25 15 
150 x  l (60 x  75 x )  21x l 2

g( x ) 
l 2  125l  150
(b) Here, f(x) = x/6, K(x, t) = 2x – t.

Proceeding as in Example 7.1(b), we obtain
2 
2 x  t  l   x  t  2 xt 
3 
R( x , t; l )  2
1  (l / 2)  (l / 6)
Hence, the required solution by Eq. (7.36) is
1
( x ) f ( x )  l  R( x, t; l ) f (t ) dt
g
0
 2 
1 2x  t  l   x  t  2 xt  
x  3 t
or g( x 
)  l   dt
6 0 1  (l / 2)  (l 2 / 6) 6
 
1
x l  2 
6 l 2  3l  6 0 
2 2
g( x )   2 x  t  l  t  xt  t  2 xt   dt
3 

1
x l  2 t3  1 2 xt 2 t 3 xt 3  
g( x )   2  xt   l  t    
6 l  3l  6  3 3 2 3 3  
0
x l  1 1 x 1 x
g( x )   2  x  l    
6 l  3l  6  3 3 2 3 3  

x l  1 lx 
g( x )   2  x 
6 l  3l  6  3 6 
1 l 
g( x
)  x 2 (6 x  2  l x ) 
6 l  3l  6 
1 l (6 x  2)  l 2 x 
g
( x) x  
6  l 2  3l  6 

7.4 FREDHOLM’S SECOND FUNDAMENTAL THEOREM

Fredholm’s first theorem does not hold good when l is a root of the equation
D(l) = 0. It has been found in Chapter 3 that for a separable kernel, the
homogeneous equation
b
g( x )  l  K ( x, t ) g(t ) dt (7.45)
a
has non-trivial solutions. It is expected that same holds when the kernel is an
arbitrary integrable function and we shall then have a spectrum of eigenvalues
and the corresponding eigenfunctions. This second Fredholm theorem is for
the study of this problem.
Statement: If l0 is a zero of multiplicity m of the function D(l), then the
homogeneous equation [Eq. (7.45)] possesses at least one, and at most m,
linearly independent solutions given by
 x1 ,..., xi 1 , x, xi 1 ,..., xr 
 gi ( x ) Dr  l0
 ; (i 1,2,..., r ,1  r  m) (7.46)
 t1 ,..., ti 1 , t , ti 1 ,..., tr 
Also, any other solution is a linear combination of these solutions. It is to be

recollected that the definition of the fredholm minor
 x1 , x2 ,..., xn   x1 , x2 ,..., xn 
Dn  l  K 
 t1 , t2 ,..., tn   t1 , t2 ,..., tn 


(l ) j
b b  x1 ,..., xn , z1 ,..., z j 
  .... K   dz1 ...dz j (7.47)
j 1 j ! a a  t1 ,..., tn , z1 ,..., z j 
where {xi} and {ti}, (i = 1,2,...,n) are two sequences of arbitrary variables,
Eq. (7.47) converges for all values of l, and hence, it is an entire function of l.
Proof: First, it will be proved that every zero of D(l) is a pole of the
resolvent kernel R(x, t; l) given by
D( x , t; l )
R( x , t; l )  (7.48)
D( l )
The order of its pole is at most equal to the order of the zero of D(l).
The Fredholm’s first series is given by

(l ) j
b b  x1 ,..., x j 
D(l ) 1    ... K   dx1 ...dx j (7.49)
j 1 j ! a a  x1 ,..., x j 
and the Fredholm's second series is given by

(l ) j
b b  x, x1 ..., x j 
, t; l ) K ( x , t )  
D( x  ... K   dx1 ...dx j (7.50)
j 1 j ! a a  t , x1 ,..., x j 
Differentiating both sides of Eq. (7.49) with respect to l and interchanging
the indices of the variables of integration, it can be expressed as
b
D1 (l )    D( x, x; l ) dx (7.51)
a
From this relation, it follows that if l0 is a zero of order p of D(l), then it is
a zero of order (p – 1) of D¢(l), and consequently, l0 may be a zero of order
at most (p – 1) of the entire function D(x, t; l). Thus, l0 is the pole of order
at most p, particularly, if l0 is a simple pole of D(l), then D(l0) = 0, D¢(l0)
≠ 0 and l0 is a simple pole of the resolvent kernel. Moreover, it follows
from Eq. (7.51) that D(x, t; l) ≠ 0. In this particular case, it is observed from
the following equation:
b
D( x, t; l ) K ( x, t ) D(l )  l  K ( x, z ) D( z, t; l ) dz

a
i.e., if D(l) = 0 and D(x, t; l) ≠ 0, then D(x, t; l), as a function of x, is a

solution of Eq. (7.45) and so is sD(x, t; l), s being an arbitrary constant.
We now consider the general case when l is a zero of an arbitrary multiplicity

m, i.e., when
D(r)(l0) = 0 and D(m)(l0) ≠ 0, (7.52)
where r = 1, 2, ..., m – 1.
Differentiating n times the Fredholm’s first series, i.e., Eq. (7.49), and obtain
b b
 x1 , x2 ,..., xn 
D( n) (l )  ( 1) n  ... K  dx1 ...dxn
 x1 , x2 ,..., xn 
a a

(l ) j
b b  x1 ,..., xn , xn 1 ,..., xn  j 
 ( 1) n    K  x1 ,..., xn , xn 1 ,..., xn  j  dx1 dx2 ...dxn  j (7.53)
...
j 1 j! a a
Now comparing Eqs. (7.51) and (7.53), we have

b b
 x1 , x2 ,..., xn 
D( n) (l )  ( 1) n  ... Dn  l  dx1 ...dxn (7.54)
a a  x1 , x2 ,..., xn 
which is a relation between nth derivative of the Fredholm function and
Fredholm minor of order n. From Eq. (7.54), it is concluded that if l0 is a zero
of order m of the function D(l), then the following holds for the Fredholm
minor of order m for that value of l0:
 x1 , x2 ,..., xm 
Dm  l0   0 (7.55)
 t1 , t2 ,..., tm 
Thus, there may exist minors of order lower than m, and which also do not
identically vanish.
We now establish the relation among the minors that corresponds to the
resolvent formula
b
, t; l ) K ( x, t )  l  K ( x, z ) R( z, t; l ) dz
R( x
a
Expanding the determinant under the integral sign in Eq. (7.47),
K ( x1 , t1 ) K ( x1 , t2 )... K ( x1 , tn ) K ( x1 , z1 )... K ( x1 , z j )
K ( x2 , t1 ) K ( x2 , t2 )... K ( x2 , tn ) K ( x2 , z1 )... K ( x2 , z j )

K ( xn , t1 ) K ( xn , t2 )... K ( xn , tn ) K ( xn , z1 )... K ( xn , z j ) (7.56)
K ( z1 , t1 ) K ( z1 , t2 )... K ( z1 , tn ) K ( z1 , z1 )... K ( z1 , z j )

K ( z j , t1 ) K ( z j , t2 )... K ( z j , tn ) K ( z j , z1 )... K ( z j , z j )
in terms of the elements of the first row and integrating j times with respect
to z1, z2, ..., zj for j  1, we obtain
b b  x1 ,..., xn , z1 ,..., z j  n
   t1 ,..., tn , z1 ,..., z j  dz1 ...dzj

... K  (1)l 1 K ( x1 , tn )
a a l 1

b b  x ,..., x ,..., x , z ,..., z  p
j  ( 1)
2 l n 1 j
  ... K  dz
 1 2dz ...dz  l  n 1
1t ,..., t , t
l 1 l 1 ,..., t ,
n 1z ,..., z j l 1
a a
b b  x2 ,..., xn , z1 , z2 ,..., zl ,..., z j 
  ... K ( x1 , zl )K   dz1 ...dz j (7.57)
a a  t1 ,..., tn 1 , tn , z1 , z2 ,..., zl 1 , zl 1 ,..., z j 
Here, it is noted that the symbols for the determinant K on the right-hand
side of Eq. (7.57) do not contain the variable xi in the upper sequence and the
variables t1 or zl in the lower sequence. Further, by transposing the variable xl
in the upper sequence to the first place by means of l + n – 2 transpositions,
it is followed that all the components of the second sum on the right side are
equal. Therefore, Eq. (7.57) can be writtern as
b b  x1 ,..., xn , z1 ,..., z j  n
  K  t1 ,..., tn , z1 ,..., z j  dz1 ...dzj

...  (1)l 1K ( x1 , tl )
a a l 1

b b  x2 ,..., xn , z1 ,..., z j 
  ... K   dz1 ...dz j
a a  t1 ,..., tl 1 , tl 1 ,..., tn , z1 ,..., z j 

b  b b  z, x2 ,..., xn , z1 ,..., z j 1  
 j  K ( x1 , z )   ... K   dz1 ...dz j 1  dz (7.58)
a  a a  t1 , t2 ,..., tn , z1 ,..., z j 1  
where the subscript l from z has been omitted.

From Eqs. (7.57) and (7.58), we find that Fredholm minor satisfies the
 x1 , x2 ,..., xn  n  x ,..., xn 
Dn  l
  (1)l 1 K ( x1 , tn )Dn 1  t ,..., t 2 
 t1 , t2 ,..., tn  l 1 1 l 1 , tl 1 ,..., t n 

b
 z, x2 ,..., xn 
 l  K ( x1 , z )Dn  l  dz (7.59)
a  t1 , t2 ,..., tn 
Expanding the determinant shown in Eq. (7.56) with respect to the first column
and proceeding as above, we obtain the integral equation:
 x1 ,..., xn  n  x ,..., x ,x ,..., xn 

Dn  l

 t1 , t ,..., tn 
 (1)l 1 K ( x1 , tn )Dn 1  t 1,.............,
l 1 l 1
t  dz
l 1 2 n

b
 x1 ,..., xn 
 l  K ( z, t1 ) Dn  dz (7.60)
a
 z, t2 ,..., tn 
The relations shown in Eqs. (7.59) and (7.60) hold for all values of l.
With the help of (7.59), one can find the solution of Eq. (7.45) for the special
case when l = l0 is an eigenvalue of the kernel, by supposing that l = l0
is a zero of multiplicity m of D(l). Then, as mentioned earlier, the minor
Dm ≠ 0 and even the minors D1, D2 ..., Dm–1 may not identically vanish.
Suppose Dr is the first minor in the sequence D1, D2 ..., Dm–1 such that Dr ≠ 0.
Then number r must be between 1 and m and is the index of eigenvalue l0.
It thus follows that Dr–1 = 0. Then Eq. (7.59) shows that
 x1 , x2 ,..., xr 
g1 ( x )  Dr  l0  (7.61)
 t1 , t2 ,..., tr 
is a solution of Eq. (7.45). Substituting x at different points of the upper
sequence in the minor Dr, we obtain r non -trivial solutions gi(x), (i = 1,2, ..., r)
of Eq. (7.45) which are often written as
 x1 ,..., xi 1 , x, xi 1 ,..., xr 
D l0 
 t1 ,................., tr 
 fi ( x )  ; (i 1,2,..., r ) (7.62)
 x1 ,..., xi 1 , xi , xi 1 ,..., xr 
Dr  l0 
 t1 ,................., tr 
in which the denominator is non–zero.
We now establish that solutions fi given by Eq. (7.62) are linearly independent.
In the determinant shown in Eq. (7.56), if we put two of the arguments xi equal,
this is equivalent to having two rows equal, and consequently, the determinant
vanishes. Thus in Eq. (7.62), we see that fk(xi) = 0 for i ≠ k, whereas fk (xk) = 1.
Now, if we have a relation of the form  ck fk  0 , then putting x = xi, we
k
get ci = 0 and so the solution fi are linearly independent. This system of
solutions fi is known as fundamental system of eigenfunctions of l0, and any
linear combination of these functions gives a solution of Eq. (7.45).
Conversely, it can be shown that any solution of Eq. (7.45) must be a
linear combination of f1(x), f2(x), ..., fv(x). For this purpose, we define a kernel
H(x, t; l), which corresponds to the resolvent kernel R(x, t; l) of Section 7.2.
 x, x1 ,..., xr 
Dr 1  l0 
 t , t1 ,..., tr 
H ( x , t; l )  (7.63)
 x1 ,..., xr 
Dr  l0 
 t1 ,..., tr 
Putting n = r and adding extra arguments x and t in Eq. (7.59), we obtain
 x, x1 ,..., xr   x1 ,..., xr  r
D
r 1  l 0 K ( x , t ) Dr 1  l0    ( 1)l K ( xl , t )
 t , t1 ,..., tr   t1 ,..., tr  l 1

b
 x, x1 ,..., xl 1 , xl 1 ,..., xr   x, x1 ,..., xr 
Dr  l0   l0  K ( z, t ) Dr 1  l0  (7.64)
 t1 , t2 ..., tr  a  z, t1 ,..., tr 
In each minor Dr, of Eq. (7.64), we transpose the variable x from the first
place to the place between the variables xl–1 and xl+1 and divide both sides
by the constant
 x1 ,..., xr 
Dr  l0   0
 t1 ,..., tr 
to get
b r
 K ( xl , t ) fl ( x ) (7.65)
H ( x, t; l )  K ( x, t )  l0  H ( x, z; l )K ( z, t )dz 
a l 1
Now suppose that g(x) is an arbitrary solution of Eq. (7.45). Multiplying both
sides of Eq. (7.65) by g(t), and then, integrating both sides with respect to t
from a to b, we get
b b r
g( x ) g( xl )
 g(t ) H ( x, t; l ) dt  
  g( z ) H ( x, z; l ) dz  fl ( x ) (7.66)
a
l0 a l 1 l 0
where we have used Eq. (7.45) in all terms except the first. Further, we have
also taken
b
 K ( xl , t )g(t )dt  g( xl )
l0
a
Cancelling the equal terms in Eq. (7.66), we obtain
r
g( x )   g( xl ) fl ( x )
l 1
It, thus, completes the proof.
7.5 FREDHOLM’S THIRD THEOREM

Statement: For the non-homogeneous integral equation of the second kind
b
g
( x ) f ( x )  l0  K ( x, t )g(t )dt (7.67)
a
to have a solution in the case D(l0) = 0, it is necessary and sufficient that the
given function f(x) is orthogonal to all the eigenfunctions hi(x), i = 1, 2, ..., n of
the transposed homogeneous equation corresponding to the eigenvalue l0, and
then, the corresponding general solution has the form
 x, x1 , x2 ,..., xr 
b
Dr 1  l 0 
 t , t1 , t2 ,..., tr  r
f ( x)  l 
g( x )  f (t )dt   clfl ( x ) (7.68)
a
 x1 , x2 ,..., xr  l 1
Dr  l0 
 t1 , t2 ,..., tr 
b
( x ) f ( x )  l  K ( x, t ) g(t )dt .
Proof: Consider g
a
Let the transpose (or adjoint) of Eq. (7.67) be

b
( x ) f ( x )  l  K (t , x ) h(t ) dt (7.69)
h
a
Now, for the transposed Eq. (7.69), Fredholm’s first and second series D(l)
and D(t, x; l) are given by
 b b  z ,..., z 
(l ) j
D ( l ) 1  
1 j

j 1
 ... K   dz ....dz j
j ! a a  z1 ,..., z j  1
(7.70)
 b b  t , z ,..., z 
(l ) j
, x; l ) K (t , x )  
1 j
and D(t
j 1
 ... K   dz ....dz j (7.71)
j ! a a  x, z1 ,..., z j  1
respectively. From this fact, it follows that kernels of Eq. (7.67)and its transpose
in Eq. (7.69) have the same eigenvalues. Further, the corresponding resolvent
kernel for Eq. (7.69) is
R(t , x; l )  D(t , x; l )/D(l ) (7.72)
and hence, the solution of Eq. (7.69) is
b
D(t , x; l )
( x) f ( x)  l 
h f (t ) dt (7.73)
a
D( l )
provided l is not an eigenvalue.
Next, it is obvious that not only the transposed kernel has the eigenvalues
as the original kernel of Eq. (7.67) but also the index r of each of the
eigenvalues is equal. Moreover, the eigenfunctions of the transposed equation
for an eigenvalue l0 are given by
 x1 ,...., xr 
Dr 
 t1 ,..., ti 1 , t , ti 1 ,..., tr 
hi (t )  (7.74)
 x1 ,...., xr 
Dr 
 t ,..., t , t , t ,..., t 
1 i 1 i i 1 r
where the values (x1, ...,xr) and (t1, ..., tr) are to be so chosen that the denominater
does not vanish. Substituting r in different places in the lower sequence of
this formula, we obtain a linearly independent system of r eigenfunctions.

Also, we know that each fi is orthogonal to each with different eigenvalues.
Suppose g(x) is a solution of Eq. (7.67). Then, multiplying Eq. (7.67) by
each member hp(x) of the above-mentioned system of functions and integrating
with respect to x from a to b, we obtain
b b bb

f ( x )h p ( x ) dx  g( x)hp ( x) dx l   K ( x, t )g(t )hp ( x)dx dt
a a aa
b b  b 
 f ( x )h p ( x ) dx    p
g ( x )  h ( x )  l  K (t, x)hp (t ) dt  dx (7.75)
a a  a 
Since hp(x) is an eigenfunction of the transposed homogeneous equation,
therefore
b
h p ( x )  l  K (t , x )h p (t ) dt (7.76)
a
where λ is the corresponding eigenvalue. Now, from Eqs. (7.75) and (7.76),
we have
b
 f ( x)hp ( x) dx  0 (7.77)
a
Thus, it follows that a necessary condition for Eq. (7.77) to have a solution
is that the non-homogeneous term f(x) be orthogonal to each solution of the
transposed homogeneous equation.
Conversely, now it will be proved that the condition Eq. (7.77) of
orthogonality is sufficent for the existence of a solution. In what follows, we
shall also obtain an explicit solution in such a case. At this stage, we define
a kernel H(t, x; l) as follows:
 x, x1 ,...., xr 
Dr 1  l0 
 t , t1 ,..., tr 
H (t , x; l )  (7.78)
 x1 ,...., xr 
Dr  l0 
 t1 ,..., tr 
wherein it is assumed that Dr ≠ 0 and that r is the index of eigenvalue l0.
To prove the required result, we show that if the orthogonality condition is
satisfied, then the function
b
( x ) f ( x )  l0  H ( x, t; l ) f (t ) dt (7.79)
g0
a
is a solution.
Substituting this value for g(x) in Eq. (7.67), we have [by referring to Eq. (7.65)]
b b  b 
f ( x )  l0  H ( x, t; l ) f (t ) dt  f ( x )  l0  K ( x, t )  f (t )  l0  H (t , z; l ) f ( z ) dz  dt
a  a 
a
b  b 
or  f (t ) dt  H (t , z; l )  K ( x, t )  l0  K ( x, z ) H ( z, t; l )dz  0 (7.80)
a  a 
Proceeding similarly, we obtain its transpose as
b r
  K ( x, t p )h p (t )
H ( x, t; l )  K ( x, t )  l0  K ( x, z ) H ( z, t; l )dz 
p 1
a
Substituting it in Eq. (7.80) and making use of the orthogonality condition,

we have an identity, and thus, it is proved.
Here, the difference of any two solutions of Eq. (7.67) is a solution of
the homogeneous equation. Hence, the most general solution of (7.67) is
b r
f ( x )  l0  H ( x, t; l ) f (t ) dt   c pf p ( x )
g( x ) 
p 1
a
EXERCISE 7.1
1. Using the Fredholm determinants, find the resolvent kernel of the

following kernels:
(a) K(x, t) = sin x – sin t, 0  x  2p, 0  t  2p
(b) K(x, t) = 2exet, a = 0, b = 1
2. Using the recursion relations, find the resolvent kernels of the
following kernels:
(a) K(x, t) = x + t + 1, –1  x  1, –1  t  1
(b) K(x, t) = sin(x + t), 0  x  2p, 0  t  2p
3. For the integral equation
b
( x ) f ( x )  l  K ( x, t ) g(t ) dt
g
a
compute D(l) and D(x, t; l) for the following kernels for the prescribed
limits a and b:
(a) K(x, t) = sin (x + t), a = 0, b = p
(b) K(x, t) = ex–t, a = 0, b = 1
4. Determine the resolvent kernel, and hence solve the following integral
equations:
1
) e x   e x t g(t ) dt
(a) g( x
0
2p
(b) 
g( x ) cos 2 x   sin x cos t g(t ) dt
0
Answers:
sin x  sin t  lp (1  2 sin x sin t )
1. (a) R( x, t; l ) 
1  4p 2 l 2
2 e x et
(b) R( x, t; l ) 
1  l (e2  1)
x + t + 1 + 2λ ( xt + 1 / 3)
2. (a) R( x, t ; λ ) =
1 − 2λ − (4 / 3)λ 2
sin( x  t )  pl cos( x  t )
(b) R( x, t; l ) 
1  p 2l2
3. (a) D( x, t ; λ ) = sin( x + t ) + ( λπ /2) cos( x − t ), D( λ ) = 1 − (π 2 /4)λ 2
(b) D( x, t; l ) x, D(l ) 1  42l
1 x
4. (a) g ( x ) = e
2
(b) g( x )  cos 2 x

Chapter 8
Integral Transform Methods
8.1 INTRODUCTION
The integral transform methods provide a useful tool for the solution of integral
equations of various special forms. Let the following double integral exist:
bb
g( x )    F ( x, t1 )K (t1 , t ) g(t ) dt dt1 (8.1)
aa
This double integral can be evaluated as an iterated integral.
b
If we take f ( x )   K ( x, t ) g(t ) dt, (8.2)
a
then from Eq. (8.1), we have

b
g( x )   F ( x, t ) f (t ) dt (8.3)
a
Thus, if Eq. (8.1) is regarded as an integral equation in g, a solution is
given by Eq. (8.2), whereas if Eq. (8.3) is regarded as an integal equation in
f, a solution is given by Eq. (8.2). It is conventional to refer, one of these
functions as the transform of the second function, and to the second function,
as an inverse transform of the first.
Some examples* are given here.
1. The most well-known double integral Eq. (8.1) is the Fourier integral
 
1
 e
 isx  ixt
g(s)  e g(t )dt dx
2p
 

which results in the reciprocal relations as:

1
e
 ist
f ( s)  g(t ) dt
2p 

* Refer to Chapter 6 of Integral Transforms published by RBD.

157

1
e
ist
and has g(s)  f (t ) dt
2p 
The function f(s) is known as Fourier transform of g(t) and g(s) as the
inverse Fourier transform of f(t), and vice versa.

2
2. Considering the double integral g(s) 
p   (sin sx sin xt ) g(t ) dx, we
0 0

2
p 0
find that this leads to the sine transform f (s)  (sin st ) g(t ) dt and its

2
p 0
inverse as, g(s)  (sin st ) f (t ) dt, respectively.
8.2 SINGULAR INTEGRAL EQUATION

Singular integral equations occur frequently in mathematical physics and
possess very unusual properties; hence, their solutions are quite essential.
Definition: An integral equation is called singular if either the range of
integration is infinite or the kernel is discontinuous.
For example, the singular integral equations of the first kind are:

f ( x )   sin ( xt ) g(t ) dt (8.4)
0

f ( x )   e  xt g(t ) dt (8.5)
0
In Eqs. (8.4) and (8.5), the range of integration is infinite, while in integral
x
g(t )
f ( x)   dt (8.6)
0 x t
the range of integration is finite, but the kernel is discontinuous.
The Abel’s Integral Equation
One of the simplest form of singular integral equation, which appears in
mechanics, is the Abel’s integral equation, given as,
x
g(t )
f ( x)  ( x  t )a dt, 0  a  1 (8.7)
0
where g(t) is an unknown function to be determined and f(x) is a known

function.
Integral Transform Methods 159
8.3 LAPLACE TRANSFORM*

Definition: Let f(x) be a function defined for x > 0. Then, the Laplace
transform of f(x), denoted by L{f(x); p} or F(p), is defined with the help of
the following integral:

e
 px
L{ f ( x
); p} F
( p) f ( x ) dx (8.8)
0
provided that the integral exists. It is to be recollected that the Laplace transform
of f(x) exists if the integral in Table 8.1 shows Laplace transform for same
elementary functions. Eq. (8.8) is convergent for some designated values of p.
Table 8.1 Laplace Transform for Some Elementary Functions
S.No. f(x) L{(x); p} or F(p)

1. 1 1/p, p > 0
n
2. x , n is positive integer n!/pn+1, p > 0
3. xn, n > –1 (n + 1)/pn+1, p > 0
4. eax 1/(p – a), p > 0
5. sin ax a/(p2 + a2), p > 0
6. cos ax p/(p2 + a2), p > 0
7. sinh ax a/(p2 – a2), p > | a |
8. cosh ax p/(p2 – a2), p > | a |
9. J0(ax)
1/ p 2 + a 2
10. Jn(ax) { p2  a2  p}n
a n p2  a2
11. d(x – a) e–ap
12. erf ( x ) 1/{ p p + 1}
–a
13. x (1 – a)pa–1
8.4 SOME IMPORTANT PROPERTIES OF LAPLACE TRANSFORM

1. Linearity property: If for i  {1,2, ..., n}, ci are constants and fi(x)
are functions with Laplace transforms Fi(p), respectively, then
L{c1 f1 ( x )   cn 
f n ( x ); p} c1 L{ f1 ( x ); p}   cn L{ f n ( x ); p}
L{c1 f1 ( x )   cn f n ( x );
p} c1F1 ( p)   cn Fn ( p)
* For details of Laplace transform and its inverse, the reader can refer to the recent book
on Integral Transforms published by RBD.
2. Change of scale property: If L{ f(x); p} = F(p), then

1
L{ f(ax); p} = F(p/a), a > 0
a
3. First shifting or translation property: If L{ f(x); p} = F(p), then
L{e–axf(x); p} = F(p + a).
4. Second shifting property: If L{ f(x); p} = F(p) and
 f ( x  a) , x  a
g( x )   , then L{g(x); p} = e–apF(p).
 0 , x  a
5. Laplace transform of derivatives: If L{ f(x); p} = F(p), then
(a) L{f '(x); p} = pF(p) – f(0),where f(x) is continuous for 0  x  N and
is of exponential order for x > N, while, f¢(x) is sectionally continuous for
0  x  N.
(b) L{f "(x); p} = p2F(p) – pf(0) – f '(0), where f(x) and f '(x) are continuous
for 0  x  N and are of exponential order for x > N, while f "(x)is sectionally
continuous for 0  x  N.
6. Laplace transform of integrals: If L{ f(x); p} = F(p), then
 x  F ( p)
L   f ( x ) dx; p  
 0  p
7. Multiplication by powers of x: If L{ f(x); p} = F(p) then
dn
L{x n f ( x ); p} 
( 1) n n F ( p) 
( 1)n F n ( p)
dp

 f ( x) 
8. Division by x: If L{ f(x); p} = F(p), then L  ; p    F (u) du,
f ( x)  x  p
provided lim exists.
x 0 x
9. Initial value theorem: Let f(x) be continuous for all x  0 and be of

exponential order as x  ¥. Also suppose that f¢(x) is of class A. Then,
lim f ( x )  lim pL{ f ( x ); p}
x 0 p
10. Final value theorem: Let f(x) be continuous for all x  0 and be of
exponential order as x  ¥. Also, suppose that f¢(x) is of class A. Then
lim f ( x )  lim pL{ f ( x ); p}
x  p 0
11. Laplace transform of periodic function: Let f(x) be a periodic

function with period a > 0, i.e., f(x + na) = f(x), for n = 1,2,...
a
1
e
 px
then, L{ f ( x ); p}  f ( x ) dx .
1  e  pa 0
8.5 INVERSE LAPLACE TRANSFORM

Definition: If the Laplace transform of a function f(x) is F(p), i.e., L{ f(x); p}
= F(p) then f(x) is called inverse Laplace transform of F(p), and we express
f ( x )  L1{F ( p); x}
where L–1 is known as the inverse Laplace transformation operator.
Table 8.2 shows inverse Laplace transform of some elementary functions.
Table 8.2 Inverse Laplace Transform* of Some Elementary Functions
S.No. F(p) L–1{F(p); x} or f(x)

1. 1/p 1
2. 1/pn+1 (n is a positive integer) xn/n!
3. 1/pa+1 {Re(a) > –1} xa/(a + 1)
4. 1/(p – a) eax
5. 1/(p2 + a2) sin ax/a
6. p/(p2 + a2) cos ax
2 2
7. 1/(p – a ) (sinh ax)/a
8. p/(p2 – a2) cos h ax
9. 1/ ( p2  a2 ) J0(ax)
10. 1/{p ( p  1)} erf ( x )
8.6 SOME IMPORTANT PROPERTIES OF INVERSE LAPLACE

TRANSFORM
1. Linearity property: If for all i  {1,2, ..., n}, ci are constants and
Fi(p) are the Laplace transforms of fi(x), respectively, then
L1{c1F1 ( p)   c 1 1
n Fn ( p); x} c1 L {F1 ( p); x}   cn L {Fn ( p); x}
L1{c1F1 ( p)   cn Fn ( p);
x} c1 f1 ( x )   cn f n ( x )
2. Change of scale property: If L–1{F(p); x} = f(x), then

1
L1
{F (ap); x} f ( x /a), a  0 .
a
3. First shifting or translation property:
If L–1{F(p); x} = f(x), then L1{F ( p  a); x} eax f ( x ) eax L1{F ( p); x}.
* The reader is advised to refer to Table 8.1 simultaneously.

4. Second shifting property: If L–1{F(p); x} = f(x) and

 0 , x  a
g( x )   ,
 f ( x  a) x  a
1  ap
then L {e F ( p); x}  g( x )
or L1{e  ap F ( p); x} f ( x  a) H ( x  a),

where H(x – a) is the Heaviside unit function.
1
5. Inverse Laplace transform of derivatives: If L {F ( p); x}  f ( x ), then
 d 
n
L1{F n ( p); x} 
L1  n F ( p); x   ( 1) n x n f ( x ), n  1, 2,
 dp 
6. Inverse Laplace transform of integrals: If L–1{F(p); x} = f(x) then
  f ( x )
L1   F (u) du; x   .
  x
 p 
7. Multiplication by powers of p: If L–1{F(p); x} = f(x) and f(0) = 0,
then L–1{pF(p); x} = f¢(x).
Further, if f (0) f  (0)  f n 1 (0) 0, then the above result is
generalised as
dn
L1{p n F ( p ); x} fn
( x) f ( x ).
dx n
8. Division by powers of p: Let L–1{F(p); x} = f(x). Then,
x
 F ( p) 
(a) L1  ; x    f (u) du,
 p  0
 F ( p)  x v
(b) L1  2 ; x     f (u) du dv
 p  0 0
 F ( p)  x x x
(c) L1  n ; x      f ( x )(dx ) n .
 p  0 0 0
8.7 CONVOLUTION OF TWO FUNCTIONS

The convolution of f(x) and g(x) is expressed and defined as
x

f *g  f (u)g( x  u) du
0
x
or f
*g  f ( x  u) g(u) du
0
The convolution theorem (or convolution property)
Let f(x) and g(x) be two functions of class A and let L–1{F(p); x} = f(x) and
L–1{G(p); x} = g(x). Then,
x
L1{F ( p) G( p); 
x}  f (u) g( x  u) 
du f * g.
0
One useful form of the above is

L{f *g) = F(p) G(p)
 x   x 
i.e., L   f (u) g( x  u) du   L   f ( x  u) g(u) du   F ( p) G( p).
 0   0 
8.8 THE HEAVISIDE EXPANSION FORMULA

Let F(p) and G(p) be polynomials in p, where F(p) has degree less than G(p).
Now, if G(p) has n distinct zeros ar, r = 1, ..., n, then
 F ( p)  n F (a r ) a r x
L1  ; x   e .
 G( p)  r 1 G '(a r )
8.9 THE COMPLEX INVERSION FORMULA

If f(x) has a continuous derivative and is of exponential order g for large
positive values of x, where g > 0 and if F(p) = L{f(x); p}, then
g  i
1
2p i g i
L1{F ( p);
 x} f
( x) e px F ( p) dp, x  0

and f(x) = 0; x < 0.
8.10 INTEGRAL EQUATIONS IN SPECIAL FORMS

1. Integral equation of convolution type: The integral equation
x
g( x )  f ( x )   K ( x  t ) g(t ) dt
0
wherein kernel K(x – t) is a function of the difference only, is known as
integral equation of convolution type. Applying the definition of convolution,
it can be expressed as
g(x) = f(x) + K(x) * g(x)
2. Integro-differential equation: An integral equation in which
derivatives of the unknown function g(x) are also present is said to be integro-
differential equation. For example,
x
d2g
 g( x )  sin x   cos( x  u) g(u) du
dx 2
0
8.11 APPLICATION OF LAPLACE TRANSFORM TO FIND THE

SOLUTIONS OF VOLTERRA INTEGRAL EQUATION
8.11.1 Convolution Type Kernels of Volterra Integral

Equation: Working Procedure
Let the Volterra integral equation of the first kind be
x

f ( x)  K ( x  t ) g(t ) dt (8.9)
0
or f ( x )  K ( x ) * g( x ) (8.10)
where kernel K(x – t) is a function of the difference (x – t).
Let L{g(x); p} = G(p), L{K(x); p} = K(p)
and L{f(x); p} = F(p) (8.11)
Now, applying the Laplace transform to both sides of Eq. (8.10) we get,
L{f(x); p} = L{K(x)*g(x)}
or F(p) = K(p) G(p) (using the convolution theroem)
or G(p) = F(p)/K(p) (8.12)
Now, applying the inverse Laplace transform to both sides of Eq. (8.12)
we get,
 F ( p) 
g( x )  L1  ; x
 K ( p) 
2. Let Volterra integral equation of the second kind be
 x
 g ( x ) = f ( x ) + ∫ K ( x − t ) g (t ) dt
g( x) =  0
(8.13)

 = f ( x ) + K ( x )* g ( x )
Applying the Laplace transform to both sides of Eq. (8.13), we get
L
{g( x ); p} L{ f ( x ); p}  L{K ( x ) * g( x )}
or G(p) = F(p) + K(p)G(p) [using Eq. (8.11) and the convolution theorem]
F ( p)
or G( p)  (8.14)
1  K ( p)
Now, applying the inverse Laplace transform to both sides of Eq. (8.14),
we obtain.
 F ( p) 
g( x )  L1  ; x
1  K ( p) 
8.11.2 Resolvent Kernel of Volterra Integral Equation by

Using Laplace Transform
In order to find the resolvent kernel of the integral equation [Eq. (8.13)] in
which kernel K(x – t) depends on the difference (x – t) by integral transform
methods, we first show that if the original kernel K(x, t) is a difference kernel,
so is the resolvent kernel.
We know that the resolvent kernel R(x, t) is a sum of its iterated kernels,
i.e.,

R( x , t )   K m ( x, t )  K1 ( x, t )  K2 ( x, t )  (8.15)
m 1
[It is to be noted that here, l = 1, so, we have used symbol R(x, t) in place
of usual symbol R(x, t; l].
Further, the iterated kernels Kn(x, t) are given by
K1(x, t) = K(x, t) (8.16)
x

and K n ( x, t ) 
K ( x, z ) K n 1 ( z, t ) dz, n 2,3,.... (8.17)
t
Since K1(x, t) = K(x, t), therefore by Eq. (8.16), we have
K1(x, t) = K(x, t) = K(x – t) (8.18)
Taking n = 2 in Eq. (8.17), we get
x x
K 2 ( x, t )   K ( x, z ) K1 ( z, t ) dz   K ( x  z ) K ( z  t ) dz
t t
x t
K 2 (
x, t )  K ( x  t  u) K (u) du [By taking z – t = u]
0
This shows that K2(x, t) depends only on the difference (x – t). Proceeding
similarly, we can show that K3(x, t), K4(x, t) ... also depend only on the
difference (x – t). Now, from Eq. (8.15), it follows that the resolvent kernel
will also depend only on the difference (x – t). Therefore, we can take that
R(x, t) = R(x – t) (8.19)
Now, it is known that the solution of Eq. (8.13) is given by
x
( x ) f ( x )   R( x, t ) f (t ) dt
g
0
x
or g( x )  f ( x )   R( x  t ) f (t ) dt (8.20)
0
Applying Laplace transform to both sides of Eq. (8.20), we get
L
{g( x ); p} L{ f ( x ); p}  L{R( x ) * f ( x )}
or G
( p) F ( p)  R( p) F ( p) (8.21)
 {g( x ); p}, F ( p) L{ f ( x ); p} and R( p)  L{R( x ); p} (8.22)
where, G( p) L
Using Eq. (8.14) for G(p) in Eq. (8.22), we get
F ( p)
 F ( p)[1  R( p)]
1  K ( p)
1 K ( p)
or 
R( p)  1 (8.23)
1  K ( p) 1  K ( p)
Now, applying the inverse Laplace transform to both sides of Eq. (8.23),
we get
 K ( p) 
L1 
R( x  t )  
1  K ( p) 
8.11.3 Solution of Integral Equations of the Type

x
f (x ) =  K (x 2 – t 2 ) g (t )dt , (x > 0) by using Laplace
0
Transform: Working Procedure
x
 K ( x  t ) g(t ) dt,
2 2
Let f ( x)  ( x  0) (8.24)
0
1 1/2 1/2 1/2
 1/2
x u , t v1/2 , g
1 ( v) v g(v ), and f1 (u)  f (u ) (8.25)
2
Then, Eq. (8.24) takes the form
u
 K (u  v) g1 (v) dv,
f1 (u)  (u  0)
0
f1 (u)  K (u) * g1 (u) (8.26)

Taking Laplace transform of both sides, we get
F1 ( p) pF1 ( p)
F1 ( p)  K ( p).G1 ( p)  G
1 ( p)  (8.27)
K ( p) pK ( p)
1
Now, let  H ( p). (8.28)
pK ( p)
Then, Eq. (8.27) becomes
 d u 
 G1 ( p) pH( p) F1 ( p) L   h (u  v) f1 (v) dv 
 du 0 
which on taking inverse Laplace transform gives
u
d
du 0

g1 (u) h (u  v) f1 (v) dv (8.29)
where h(x) = L–1{H(p); x}.

Finally, from Eq. (8.25) and (8.29), we have the required solution as
x
d
dx 0
g( x ) 2 t f (t ) h( x 2  t 2 )dt (8.30)
EXAMPLE 8.1: Solve the following Abel’s integral equation:

x
g(t )

(a) f ( x )  ( x  t )a dt, 0 a 1
0
x
g(t )
 ( x  t )1/2 dt 1  x  x
2
(b)
0
Solution: (i) The given integral equation is of convolution type, and therefore
expressing it as:
f(x) = g(x)*x–a (i)
Taking Laplace transform of both sides of Eq. (i) and applying the
convolution theorem, we have
L{ f ( x ); p}  L{g( x ); p}. L{x a ; p}
(1  a )
or F ( p)  G( p)
p1a
p1a F ( P) p
or 
G( p)  [ (a ) p a F ( p)]
(1  a ) (a )(1  a )
p
G ( p) [ (a ) p a F ( p)]
(p / sin pa )
(∵ (a ) (1  a ) p / sin pa )
p sin pa
G( p)  L{x a 1 * f ( x )},
p
sin pa  x 

G( p) pL   ( x  t )a 1 f (t ) dt  (By convolution theorem) (ii)
p  0 
x
Let h
( x)  ( x  t )a 1 f (t ) dt (iii)
0
Now, L h '( x
); p pL {h( x ); p} 
h(0) pL{( x ); p} {∵h(0)  0}
 x 
or pL   ( x  t )a 1 f (t ) dt; p   L{h '( x ); p} [by Eq. (iii)] (iv)
 0 
Using Eq. (iv) in Eq. (ii), we get
sin pa
G( p) 
L{h '( x ); p}
p
Now, taking inverse, we have
x 
sin pa d  ( x  t )a 1 f (t ) dt 
 g( x ) L 1
{G( p); x} 
p dx  0 
(b) Rewriting the given equation in convolution form, we have

g( x ) * x 1/2 1  x  x 2 (v)
Taking Laplace transform of both sides of Eq. (v) and using the convolution
theorem,we have
L{g( x )}  L{x 1/2 }  L (1)  L{x}  L{x 2 }
(1 / 2) 1 1 2!
or G( p) 1/2
  2  3
p p p p
1  1 1 2 
or G
( p)  1/2  3/2  5/2  (vi)
p p p p
Now, applying the inverse Laplace transform to both sides of Eq. (vi),we
obtain
1  1  1  1  1   1 
1 
g( x)  L  1/2   L  3/2   2 L  5/2 
p   p   p   p 

1  x 1/2 x1/2 2 x 3/2 
g
( x)    
p  (1 / 2) (3 / 2) (5 / 2) 

1 1/2
g( x
) [x  2 x1/2  (8 / 3) x 3/2 ], (Upon simplification)
p
x
EXAMPLE 8.2: Solve the integral equation 
sin x  J0 ( x  t ) g(t ) dt.
0
Solution: The given integral equation may be expressed in convolution form
as
sin
x = g(x)*J0(x) (i)
Taking Laplace transform of both sides of Eq. (i) and using the convolution
theorem, we have
L{sin x}  L{g( x )} L{J 0 ( x )}
1 1
or 2
 G( p)
p 1 p 12
1
or G( p) =
p2 + 1
Now, taking the inverse Laplace transform, we have
1
g( x ) L {G( p); x} J 0 ( x )
x
EXAMPLE 8.3: Solve the integral equation x   e
x t
g(t ) dt .
0
Solution: The integral equation may be written in convolution form as
x = ex * g(x) (1)
Now, taking Laplace transform of both sides of Eq. (i) and using the
convolution theorem, we get

L{x} L{e x * g
( x )} L{e x } L{g( x )}
1 1
or  G( p)
p 2 p 1
p 1 1 1
or G( p
)   (ii)
p2 p p2
Taking the inverse Laplace transform of Eq. (ii),we obtain

1   1 
g( x ) L1{G
 ( p); x} L1    L1  2 
 p  p 
or g(x) = 1 – x.
x
1   sin( x  t ) g(t ) dt and
EXAMPLE 8.4: Solve the integral equation g( x ) 
verify your answer. 0
Solution: The given integral equation can be rewritten in convolution form as

g(x) = 1 + g(x)*sin x (i)
Taking Laplace transform of both sides of Eq. (i) and using the convolution
theorem, we get
L{g(
x )} L{1}  L{g( x )} L{sin x}
1 1
or G( p
)  G ( p) 2
p p 1
 1  1
or 1  2  G ( p) 
 p  1 p
p2  1 1 1
or G( p
)  
p3 p p3
Applying Laplace inversion, we get
x2 x 2 (ii)
g( x ) 

1 1
2! 2
Verification of solution: N
ow, we show that Eq. (ii) satisfies the given
integral equation.
(iii)
From Eq. (ii), g(x) = 1 + (x2/2)
Thus, the R.H.S. of Eq. (iii) provides
x
1   sin ( x  t )(1  t 2 /2) dt
0
x
 t 2   x
1  1   cos( x  t )    t cos( x  t ) dt

  2 0 0

x2  x 
 11  cos x  [ t sin( x  t )]0   sin ( x  t ) dt 
x
2  0 

x2
2   cos x  [cos( x  t )]0x
2
x2 x2
2   cos x  (1  cos x ) 1  g( x );
2 2
= L.H.S. of Eq. (iii)
EXAMPLE 8.5: Solve the following Volterra integral equation of the first kind:
x
 g(t )g( x  t )dt 

16 sin 4 x
0
Solution: The given integral equation can be written as
g(x)*g(x) = 16 sin 4x (i)
Taking Laplace transform to both sides of Eq. (i) and applying the
 4 
L{g( x )}2  16  2
 p  16 

8
or L{g( x )}  
2
p  16
Now, taking inverse Laplace transform,we have
  8 
g( x )  L1     8 J 0 (4 x )
2
 p  16 
EXAMPLE 8.6: Solve the following integral equations:

x
e  x 2  cos( x  t ) g(t ) dt
(a) g( x ) 
0
x
1
x   ( x  t )3 g(t ) dt
(b) g( x ) 
6 0
Solution: (a) Rewriting the given integral equation, we have

g(x) = e–x – 2g(x)*cos x (i)
Applying Laplace transform to both sides of Eq. (i) and using the
g( x )} L{e  x }  2 L [ g( x )] L{cos x}
L{
1 p
or G
( p)  2G( p) 2
p 1 p 1
p2  1
or G( p) 
( p  1)3
[( p  1)  1]2  1
or G( p)  (ii)
( p  1)3
Now, applying inversion of Eq. (ii), we get
 [( p  1)  1]2  1 
g( x )  L1  
 ( p  1)3 

 ( p  1)  1 
2
g( x )  e  x L1   (By first shifting theorem)
 p3 
 p2  2 p  2   x 1  1 2 2 
g( x ) e  x L1 
 3
  e L   2  3
 p   p p p 
 1   1   1 
g( x ) e  x  L1    2 L1  2   2 L1  3 
  p  p   p 
g( x ) e  x [1  2 x  2( x 2 / 2!)]
g( x ) e  x (1  2 x  x 2 )
or g( x ) e  x (1  x )2

(b) Rewriting the given integral equation, we have
1 3
g( x ) x 
x * g( x ) (iii)
6
Applying Laplace transform to both sides of Eq. (iii) and using the
1
x )} L{x}  L{x 3}. L{g( x )}
L{g(
6
1 1 3!
or G( 
p)  G( p)
p2 6 p 4
 1  1
or G( p) 1  4  
 p  p2
1 1 1 
or 
G ( p)  2  2 
2  p  1 p  1
Now, applying inversion, we obtain
1

g( x ) (sinh x  sin x ).
2
EXAMPLE 8.7: Solve the following integro-differential equations
x
'( x ) sin x   g(t ) cos t dt , where g(0) = 0.
(a) g
0
x
(b) g '( x )  5 cos 2( x  t ) g(t )dt 
10, where g(0) = 2.
0
Solution: (a) The given integral equation can be written as
g'(x) = sin x + g(x)*cos x (i)
Also, given that g(0) = 0. (ii)
convolution theorem, we obtain
L
{g '( x )} L{sin x}  L{g( x )}L{cos x}
1 p
or p G( p)  g(0) 2
 G( p) 2
p 1 p 1
 1  1
or 1   p G( p)  [Using Eq. (ii)]
2 2
 p  1 p 1
1
or G( p)  (iii)
p3
Now, inverting Eq. (iii) to get
 1  x
2
x2
x ) L1  3
g(   .
 p  2! 2

(b) The given integral equation may be written as
g'(x) + 5[cos2x*g(x)] = 10 (iv)
Also, given that g(0) = 2. (v)
Applying Laplace transform to both sides of Eq. (iv) and using the
L{g '( x )}  5L {cos 2 x} L{g( x )} 

10 L {1}
p 10
or p G( p)  g(0)  5 2
G( p) 
p 4 p
 5  10
or 1  2  pG( p) p
 2, [Using Eq. (v)]
p 4
2 p  10 p2  4 2 p3  10 p2  8 p  40
or G( p)   2 
p2 p 9 p2 ( p2  9)
or 8 40 10 p  50 (vi)
G( p)   2 
9p 9p 9( p2  9)
Now, taking inverse Laplace transform of Eq. (vi), we have
 8 40 10 p  50 
) L1   2 
g( x 
 9 p 9 p 9( p2  9) 

8  1  40 1  1  10 1  p  50 1  1 
g( x ) L1    L  2 L  2  L  2 

9  p 9  p  9  p  9  9  p  9 
8 40 10 50
g( x )  x  cos3 x  sin 3 x
9 9 9 27
1
or g( x )  (24  120 x  30 cos3 x  50 sin 3 x )
27
EXAMPLE 8.8: Find the resolvent kernel of the following Volterra integral
equations, and hence, find their solutions:
x
(a) g( x )  f ( x )   ( x  t ) g(t ) dt
0
x
(b) g ( x ) f ( x )   e( x t ) g(t ) dt
0
Solution: (a) The given integral equation can be written as
g(x) = f(x) + g(x)*x (i)

L ( g( x )  L{ f ( x )}  L{g( x )}  L{x}

or 1
G( p) F ( p)  G( p) 2
p
 1 
or
1  2  G( p)  F ( p)
p
p2
or G ( p)  2 F ( p) (ii)
p 1
Let R(x – t) be the resolvent kernel of the given integral equation. The
required solution is given by
x
g( x )  f ( x )   R( x  t ) f (t ) dt (iii)
0
or g
( x ) f ( x )  R( x ) * f ( x ) (iv)
Applying Laplace transform to both sides of Eq. (iv) and using the
L{g( x )}  L{ f ( x )}  L{R( x )} L{ f ( x )}
or ( p) F ( p)  R( p) F ( p) [where R( p)  L{R( x )} ]
G
p2
or F
( p) F ( p)  R( p) F ( p) [Using Eq. (ii)]
p2  1
p2 1
or R
( p) 2
1 2
p 1 p 1
Now, inverting, R(x) = L–1{R(p)} = sinh x
so that R(x – t) = sinh (x – t)
giving the required resolvent kernel.
Now, substituting the above value of R(x – t), the required solution by
Eq. (iii) is
x
g( x ) f ( x )   sinh ( x  t ) f (t ) dt
0
(b) The given integral equation can be written as
g(x) = f(x) + g(x)*ex (v)
Applying Laplace transform to both sides of Eq. (v) and using the
{g( x )} L{ f ( x )}  L{g( x )}. L{e x }
L
1
or G( p)  F ( p)  G( p)  G( p)
p 1
 1 
or 1  G ( p) 
F ( p)
 p  1 
p 1
or G( p)  F ( p) (vi)
p2
Let R(x – t) be the resolvent kernel of the given integral equation. Then,
we know that the required solution is given by
x
g( x )  f ( x )   R( x  t ) f (t )dt (vii)
0
or g
( x ) f ( x )  R( x ) * f ( x ) (viii)
Applying Laplace transform to both sides of Eq. (viii) and using the
L
{g( x )} L{ f ( x )}  L{R( x )}. L{ f ( x )}
or G
( p) F ( p)  R( p) F ( p), where R( p)  L{R( x )}
p 1
or F ( p) F ( p)[1  R( p)], [Using Eq. (vi) for G(p)]
p2
p 1 1
or R(
p) 
 1
p2 p2
1  1  2x

Now inverting, R( x ) L   e
 p  2 
Hence, e2( x t )
R( x  t ) 
It is the required resolvent kernel.
And now, substituting the value of R(x – t), the required solution by Eq. (vii) is
x
( x ) f ( x )   e2( x t ) f (t ) dt
g
0
EXAMPLE 8.9: Solve the following integral equation

x
g(t )
 f ( x)  ( x 2  t 2 )a dt (0  a  1)
0
Solution: Substituting [refer to Eqs. (8.24) and (8.25)]
1 1/2 1/2

1/2
x u , t n 1/2 , g
1 (n ) n g(n ), and f1 (u)  f (u1/2 ) (i)
2
Then, the given integral equation takes the form
u
g1 (u )
f1 (u ) = ∫ dv (ii)
0 (u − v )α
Now, taking Laplace transform of Eq. (ii), we get
p) G1 ( p)(1  a ) pa 1 [since L{u a ; p} (1  a ) pa 1 ]

F1 ( 
pF1 ( p)
Thus, G1 ( p)  a
p (1  a )
1
and H ( p)  a
[Refer to Eqs. (8.27) and (8.28)]
p (1  a )
Then, h( x )  L1{H ( p); x}

1  1  sin ap a 1

h(x) L  a ; x x
 p   a )  p
Thus, solution of the given integral equation by (i) is

x
2sin ap d t f (t )
g( x ) 
p 
dx 0 ( x  t 2 )1a
2
. dt

8.12 FOURIER TRANSFORMS AND THEIR IMPORTANT

PROPERTIES*
1. The Fourier or complex Fourier transform: Let f(x) be a function
defined on (–¥, ¥) and be piecewise continuously differentiable and absolutely
integrable in (–¥, ¥). Then, the Fourier transform of f(x), denoted by F{f(x); p}
or F(p), is defined as

1
e
 ipx
F ({ f ( x
); p} F
( p) f ( x )dx, (–¥, < p < ¥) (8.31)
2p 
1
Here, e ipx is known as the kernel of the Fourier transformation and
2p
F represents Fourier transformation operator.
Then, the function f(x) is called inverse Fourier transform of F(p) and is
given by

1
e
ipx
f ( x)  F ( p) dp (8.32)
2p 
2. The (infinite) Fourier sine transform: Let f(x) be a function defined

for x > 0 and be piecewise continuously differentiable and absolutely integrable
in (0, ). Then, the Fourier sine transform of f(x), denoted by Fs{f(x); p}or
Fs(p), is defined as

2
Fs { f ( x
); p} F
s ( p)
p  f ( x)sin ( px) dx,( p  0) (8.33)
0
2
Here, sin px is known as the kernel of the Fourier sine transform
p
and Fs represents the Fourier sine transformation operation.
The corresponding inversion formula is then given by

2
f ( x) 
p  Fs ( p)sin ( px) dp (8.34)
0
3. The (infinite) Fourier cosine transform: Let f(x) be a function
defined for x > 0 and be piecewise continuously differentiable and absolutely
integrable in (0, ¥). Then, the Fourier cosine transform of f(x), denoted by
Fc{f(x); p} or Fc(p), is defined as

2
Fc { f ( x
); p} F
c ( p)
p  f ( x) cos( px) dx, ( p  0) (8.35)
0
2
Here, cos( px ) is the kernel of the Fourier cosine transform and Fc
x
represents the Fourier cosine transformation operation.
* For more details, readers can consult Integral Transforms by RBD.

The corresponding inversion formula is given by


2
f ( x) 
p  Fc ( p) cos( px) dp (8.36)
0
Remarks: (a) If the kernel of the Fourier transform is taken as eipx, then
Eqs. (8.31) and (8.32) take the following forms:
 
1
e e
ipx  ipx
F ( p)  f ( x ) dx and f ( x )  F ( p)dp

2p 
(b) If the kernel of the Fourier sine transform is taken as sin px, then
Eqs. (8.33) and (8.34) take the forms as:
 
2
Fs ( p) 
p 0
 f ( x )sin ( px ) dx and f ( x ) Fs ( p)sin px dp
0
(c) If the kernel of the Fourier cosine transform is taken as cos px, then
Eqs. (8.35) and (8.36) take the forms as
 
2
 Fc ( p) 
f ( x ) cos( px ) dx and f ( x )
p  Fc ( p) cos( px) dp
0 0
4. Linearity of Fourier transforms: Let ci be constants and fi(x) [for all

i  {1,2,..., n}] be functions with Fourier transforms Fi(p), respectively, then
F{c1 f1 ( x )   cn
f n ( x ); p} c1F{ f1 ( x ); p}   cn F{ f n ( x ); p}
 c1F1 ( p)   cn Fn ( p)
This property holds good for sine and cosine transforms also.
5. Change of scale property:
1

(a) If F{ f ( x ); p} F
( p), then F{ f (ax ); p} F ( p /a) .
a
1

(b) If Fs { f ( x ); p} Fs ( p), then Fs { f ( ax ); p} Fs ( p /a) .
a
1

(c) If Fc { f ( x ); p} F
c ( p), then Fc { f ( ax ); p} Fc ( p /a) .
a
6. Shifting property:
If F x  a); p} eipa F ( p).
{ f ( x ); p} F ( p), then F{ f (
7. Convolution or faltung of two functions: The convolution of two
integrable functions f(x) and g(x), where –¥, < x < ¥, is expressed and defined
as

1
f *g
2p
 f (u) g( x  u) du

8. The convolution theorem for Fourier transforms–Statement: Let
(a) f(x) and g(x) and their first order derivatives be continuous on (¥, ¥),
(b) f(x) and g(x) be absolutely integrable on (¥, ¥),
(c) F(p) and G(p) be Fouier transforms of f(x) and g(x); respectively.
Then, the Fourier transform of the convolution of f(x) and g(x) exists and is
the product of the Fourier transforms of f(x) and g(x), i.e.,
F{ f * g}  F{ f ( x ); p} . F{g( x ); p}
8.13 APPLICATION OF FOURIER TRANSFORM TO DETERMINE

THE SOLUTION OF SINGULAR INTEGRAL EQUATIONS
The procedure will be made clear through the following examples:
EXAMPLE 8.10: Solve the integral equation for f(x)


1  p , 0  p  1
 f ( x) cos px dx   0 , p  1
0

sin 2 t p
Hence, deduce that  t 2
dt 
2
.
0

2
Solution: Let
p  f ( x) cos
 px dx F
c { f ( x )} Fc ( p).
0
 2
 (1  p), 0  p  1
Then, Fc ( p)   p
 0, p 1

Hence, by the Fourier cosine inversion formula,we have
 1
2 2 2
 f ( x)
p  Fc ( p
) cos px dp
p  p
(1  p) cos px dp
0 0
2(1  cos x )
f(x) 
p x2
This is the required solution.
Deduction: Substituting the value of f(x) in the given integral equation,

we get

2(1  cos x ) 1  p, 0  p  1
 p x 2 cos px dx   0, p 1
(i)
0
Taking p  0, Eq. (i) yields

 
2 1  cos x 2 sin 2 ( x / 2) p
 
p0 x 2
dx 1or  x 2
dx
2
0
Putting x = 2t, dx = 2dt, we get


sin 2 t p
 t 2
dt  .
2
0
EXAMPLES 8.11: Solve the following integral equation for f(x):
 1 , 0  p  1

 f ( x )sin px 
dx 2 , 1  p  2
0 0 , p  2


2
Solution: Let
p  f ( x)sin
 px dx F
s { f ( x )} Fs ( p).
0
1 , 0  p  1
2
Then, 
Fs ( p) 2 , 1  p  2
p
1 , p  2
Hence, by the Fourier sine inversion formula, we get

2
f ( x) 
x  Fs ( p)sin px dp
0
1 2 
2 2 2
p 0
f ( x)  1.sin px dp   2sin px dp   0sin px dp
p1 p2

1 2
2   cos px  4   cos px 
f ( x)     
p x 0 p  x 1

2 4
f ( x ) [  cos x  1]  { cos 2 x  cos x}
px px
2
f ( x)  (1  cos x  2 cos 2 x )
px
EXERCISE 8.1
1. Show that the solution of the integral equation

x
g(t ) dt 3 3
 ( x  t )1/3 
(1/3)
x (1  x ) is x (3 x  2) .
0
4p
2. Solve the following Abel’s integral equation
x
1
 ( x  t )1/2 g(t )dt  t
0
3. Solve the following inhomogeneous integral equations:

x
x 2  sin( x  t ) g(t ) dt
(a) g( x ) 
0
x
1   ( x  t ) g(t ) dt
(b) g( x ) 
0
Also, verify the results so obtained.
4. Solve the following integral equations:
x
c sin x  2  cos( x  t ) g(t ) dt
(a) g( x ) 
0
x
x  2  cos( x  t ) g(t ) dt
(b) g( x ) 
0
x
(c)  g(t ) g( x  t )
dt 2 g( x )  x  2
0
x
(d)  g(t ) g( x  t ) dt 
x 2 g( x )
0
5. Show that the only solution of the integral equation
x
 g(t )sin ( x  t ) dt 
g( x )
0
is the trivial solution g(x) = 0.

6. Solve the following equations:
x
(a)  g(t ) cos(
x  t ) dt g '( x ) 
if g(0) 1
0
x
(b)  g '(t ) g( 
x  t ) dt 24 x 3 
if g(0) 0
0
x
(c)  g ''(t )g '( x  t ) dt  g '( x )  g( x ) if g(0)  g '(0)  0
0
x
x   g( x  t )cos t dt , if g(0) 
(d) g '( x )  4
0
7. Solve the following integral equation:

x
 f ( x) cos px dx  e
p
8. Find resolvent kernel of the following integral equation:

x
g( x ) f ( x )  l  J 0 ( x  t ) g(t ) dt
0
9. Solve the following inhomogeneous Abel’s integral equation:

x
g(t )
) f ( x)  l 
g( x dt (0  a  1)
( x  t )a
0
10. Solve the integro-differential equation

x
g ''( x )   e2( x t ) g '(t )dt 
e2 x
0
if 
g(0) g
'(0) 0
Answers
1
g(t ) 
2.
2
x2
) x2 
3. (a) g( x (b) g(x) = cos x
12
x
4. (a) g( x )  cxe
(b) g( x )  x  2  2( x  1)e x
(c) g(x) = 1
x x
(d) g( x )  J1 ( x )   J 0 (t )dt or g( x )  2d ( x )  J1 ( x )   J 0 (t ) dt
0 0
x2 16 x 3/2
6. (a) g( x ) 1  (b) g ( x ) = ±
2 π
(c) g(x) = 0,
5 2 1 4
(d) g ( x ) = 4 + x + x
2 24
2
7. f ( x ) =
π (1 + x 2 )
∫ (sin ) ( )
x
λ J1 (t )
8. 1− λ2 (x − t) dt + λ cos x 1 − λ 2
(1 − λ )2 0
t

+
λ2
1− λ 2 (
sin x (1 − λ 2 )
x
g ( x ) = f ( x ) + ∫ R( x, t ) f (t ) dt
9.
0
∞
1
where, R( x, t ) = ∑
( x − t )Γ[n(1 − α )] n =1
[λΓ (1 − α )( x − t )1−α ]n
g ( x ) = xe x − e x + 1
10.


Index
Abel’s integral equation, 2, 158 Difference kernel, 4

Abel’s problem, 1
Application of Fourier transform, 179
Application of Laplace transform, 164 Eigenfunction, 7, 30
Approximate eigenvalues, 123 Eigenvalues, 7, 29
Entire function, 125
Exponential order, 160
Bessel’s inequality, 62
Bivariate function, 4
Boundary conditions, 1 Final value theorem, 160
Boundary value problem, 1, 3, 21 Fourier coefficients, 61
Fourier cosine transform, 177
Fourier cosine transformation operation,
Change of scale property, 160, 178 177
Characteric numbers, 7 Fourier integral, 157
Characteristic function, 7, 29 Fourier or complex Fourier transform,
Characteristic number, 29 177
Characteristic value, 7, 29 Fourier sine transform, 177
Complex Fourier transform, 177 Fourier sine transformation operation, 177
Complex inversion formula, 163 Fourier transform, 158, 177
Convolution, 162 Fourier transformation operator, 177
integral equation, 5, 6 Fredhlom’s first series, 125
of two functions, 162 Fredholm determinant, 122, 129
or faltung of two functions, 178 Fredholm integral equation, 2, 5
property, 163 first kind, 5
Convolution theorem, 163 second kind, 5
for Fourier transform, 179 third kind, 5
Fredholm minor, 129
Fredholm resolvent kernel, 128
Degenerate kernel, 4, 14 Fredholm’s determinant, 125
Denumerable, 59 Fredholm’s first fundamental theorems,
Determination of resolvent kernel, 96 121
185
186 Index
Fredholm’s second fundamental theorem, of integrals, 160

147 of periodic function, 160
Fredholm’s second series, 126, 148 Linear equations, 3
Fredholm’s third theorem, 152 Linear integral equation, 4
Fundamental functions, 7 Linearity of Fourier transforms, 178
Fundamental properties of Linearity property, 159
eigenvalues of symmetric kernels, 59
eigenfunctions of symmetric kernels, 59
Fundamental system of eigenfunctions of Meromorphic function, 121
l0, 151 Method of
conversion of a boundary value
problem to a Fredholm integral
Heaviside expansion formula, 163 equation, 21
Hilbert–Schmidt theorem, 60 conversion of an initial value problem
Homogeneous Fredholm integral to a Volterra integral equation, 12
equations, 29 transforming the initial value problem
Homogeneous Volterra integral equation into a Volterra integral equation, 18
of second kind, 6
Neumann series, 89, 96

Index of eigenvalue, 151 Non-linear integral equation, 3
Initial value problem, 1, 2 Normalised eigenfunctions, 63
Initial value theorem, 160 Norm of a function, 58
Inner or scalar product of two functions, 58
Integrable function, 125
Integral equation, 1, 3 Operator adjoint, 58
of convolution type, 6, 163 Orthogonality of
in special forms, 163 eigenfunctions, 33
Integro-differential equation, 3, 163 two functions, 33
Inverse Fourier transform, 158, 177 Orthogonal system of functions, 59
Inverse Laplace transform, 161 Orthonormal, 59
Inverse Laplace transformation operator,
161
Iterated kernel, 6, 84 Proper orthogonal set, 59
Iterative method, 89 Properties of inverse Laplace transform,
Iterative scheme, 89 161
Properties of Laplace transform, 159
Kernel, 3
of Fourier cosine transform, 177 Real eigenvalues, 34
of Fourier sine transform, 177, 178 Reciprocal functions, 104
Kernel of Fourier transform, 178 Reciprocal kernal, 7, 85
Kernel of Fourier transformation, 177 Regularity condition, 57
Kinds of linear integral equations, 4 Resolvent determinant, 123
Resolvent kernel, 7, 84, 85
Laplace transform, 159

of derivatives, 160 Scalar, 58
Index 187
Schmidt’s solution, 63 Square integrable function, 58
Second shifting property, 160 Symmetric kernel, 4
Sectionally continuous, 160
Self-adjoint, 58
Separable or degenerate kernel, 4 Transposed kernel, 153
Shifting property, 178 Trivial solution, 29
Singular integral equation, 5, 6, 158
Solution of an integral equation, 8
Solution of Fredholm integral equation, Volterra integral equation, 2, 5
98, 102 of first kind, 5
Solution of Volterra integral equation, of second kind, 6
111, 113 of third kind, 5
Special kinds of kernels, 4
Spectrum, 7
of eigenvalues, 147 Zero or trivial solution, 7

4 Integral Equations by D.C. Sharma M.C. Goyal

Uploaded by

Document Information

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

4 Integral Equations by D.C. Sharma M.C. Goyal

Uploaded by

Copyright:

Available Formats

Integral Equations

2. Applications to Ordinary Differential Equations....... 12–28

3. Solution of Homogeneous Fredholm Integral

4. Fredholm Integral Equations with

5. Integral Equations with Symmetric Kernels............... 56–83

6. Solution of Integral Equations of the Second

7. Classical Fredholm Theory..................................... 121–156

8. Integral Transform Methods.................................. 157–183

8.11 Application of Laplace Transform to Find the Solutions of Volterra

Transform: Working Procedure 166

Integral Equations are being used as an essential effective tool by the

1.2 ABEL’S PROBLEM

1.3 INITIAL VALUE PROBLEM AND BOUNDARY VALUE PROBLEM

Initial value problem

1.4 INTEGRAL EQUATION

is integro-differential equation of unknown function u(x).

The most general linear integral equation may be shaped as

2. Second kind, if v = 1, Eq. (1.2) provides

3. Third kind, if v  0 , Eq. (1.2) itself works.

1.5 SPECIAL KINDS OF KERNELS

1.6 CLASSIFICATION OF INTEGRAL EQUATION

and it is Fredholm integral equation of the first kind.

it is known as homogeneous Fredholm integral equation of the second kind.

respectively, and are called integral equations of convolution type.

1.7 ITERATED KERNELS

1.8 RECIPROCAL KERNAL OR RESOLVENT KERNEL

1.9 EIGENVALUES AND EIGENFUNCTIONS

for which u(x) = 0 is an obvious solution. This solution is taken as zero or

1.10 SOLUTION OF AN INTEGRAL EQUATION

the solution is u(x) = ex, which is verified immediately.

(Now, integrating by parts)

L.H.S. of the integral equation

px p2    cos(px /2)  x x  cos(px /2)  

2.2 METHOD OF CONVERSION OF AN INITIAL VALUE

Integrating Eq. (2.4) with respect to x from a to x, we obtain

Proceeding in this way, we get it reduced to

EXAMPLE 2.1: Transform the initial value problem

which on using Eq. [ii(b)], becomes

which is the required equation.

EXAMPLE 2.2: Transform the initial value problem

Again, integrating Eq. (iii) with respect to x from 0 to x, we obtain

EXAMPLE 2.3: Form an integral equation corresponding to the differential

Integrating Eq. (iii) again with respect to x from 0 to x, we get

1. Reduce the following initial value problems to Volterra integral equations

y ′′′ + xy ′′ + ( x 2 − x ) y = xe x + 1 , y(0) = 1, y¢(0) = 1, y¢¢(0) = 0

2.3 ALTERNATE METHOD OF TRANSFORMING THE INITIAL

EXAMPLE 2.4: Transform the following initial value problem into an

Integrating again with respect to x from a to x,

This is the required integral equation.

Integrating Eq. (iii) with respect to x from 0 to x, we get

For this, we apply Leibnitz’s rule of differentiation under the sign of

Differentiating Eq. (v) with respect to x

1. Convert the following initial value problems into integral equations.

2.4 BOUNDARY VALUE PROBLEM AND ITS CONVERSION TO

EXAMPLE 2.6: Reduce the following boundary value problem into an

Now using Eq. [i(a)],

Now, Eq. (ii) can be expressed as

EXAMPLE 2.7: Transform the boundary value problem

Let y¢(0) = c, so that

Integrating both sides with respect to x

Using Eq. [i(a)] we express Eq. (iii) as

Now, to determine c, we differentiate [iv(a)] and use Eq. [i(b)].

Differentiating both sides with respect to x, we have

Again, differentiating with respect to x, we get

or c1  b1  l[c1a11  c2a12   cna1n ]