MM-165 CH001.

tex 26/4/2007 13: 43 Page 5

Introduction 5

was only by manipulation, we obtained the integral equation. In the Abel’s problem,
the physics gave us the integral equation directly. In any event, observe that we
can solve the integral equation by application of the Laplace Transform. Integral
equations of the convolution type can easily be solved by the Laplace transform .

1.4 Classification of integral equations

An integral equation can be classified as a linear or nonlinear integral equation as we
have seen in the ordinary and partial differential equations. In the previous section,
we have noticed that the differential equation can be equivalently represented by
the integral equation. Therefore, there is a good relationship between these two
equations.
The most frequently used integral equations fall under two major classes, namely
Volterra and Fredholm integral equations. Of course, we have to classify them as
homogeneous or nonhomogeneous; and also linear or nonlinear. In some practical
problems, we come across singular equations also.
In this text, we shall distinguish four major types of integral equations – the
two main classes and two related types of integral equations. In particular, the four
types are given below:

• Volterra integral equations

• Fredholm integral equations
• Integro-differential equations
• Singular integral equations

We shall outline these equations using basic definitions and properties of each type.

1.4.1 Volterra integral equations

The most standard form of Volterra linear integral equations is of the form

x
φ(x)u(x) = f (x) + λ K(x, t)u(t)dt (1.8)
a

where the limits of integration are function of x and the unknown function u(x)
appears linearly under the integral sign. If the function φ(x) = 1, then equation
(1.8) simply becomes

x
u(x) = f (x) + λ K(x, t)u(t)dt (1.9)
a

and this equation is known as the Volterra integral equation of the second kind;
whereas if φ(x) = 0, then equation (1.8) becomes

x
f (x) + λ K(x, t)u(t)dt = 0 (1.10)
a

which is known as the Volterra equation of the first kind.

MM-165 CH001.tex 26/4/2007 13: 43 Page 6

6 Integral Equations and their Applications

1.4.2 Fredholm integral equations

The most standard form of Fredholm linear integral equations is given by the form

b
φ(x)u(x) = f (x) + λ K(x, t)u(t)dt (1.11)
a

where the limits of integration a and b are constants and the unknown function
u(x) appears linearly under the integral sign. If the function φ(x) = 1, then (1.11)
becomes simply

b
u(x) = f (x) + λ K(x, t)u(t)dt (1.12)
a

and this equation is called Fredholm integral equation of second kind; whereas if
φ(x) = 0, then (1.11) yields

b
f (x) + λ K(x, t)u(t)dt = 0 (1.13)
a

which is called Fredholm integral equation of the first kind.

Remark
It is important to note that integral equations arise in engineering, physics, chem-
istry, and biological problems. Many initial and boundary value problems associated
with the ordinary and partial differential equations can be cast into the integral
equations of Volterra and Fredholm types, respectively.
If the unknown function u(x) appearing under the integral sign is given in
the functional form F(u(x)) such as the power of u(x) is no longer unity, e.g.
F(u(x)) = un (x), n  = 1, or sin u(x) etc., then theVolterra and Fredholm integral equa-
tions are classified as nonlinear integral equations. As for examples, the following
integral equations are nonlinear integral equations:

x
u(x) = f (x) + λ K(x, t) u2 (t) dt
a

x
u(x) = f (x) + λ K(x, t) sin (u(t)) dt
a

x
u(x) = f (x) + λ K(x, t) ln (u(t)) dt
a

Next, if we set f (x) = 0, in Volterra or Fredholm integral equations, then the result-
ing equation is called a homogeneous integral equation, otherwise it is called
nonhomogeneous integral equation.
MM-165 CH001.tex 26/4/2007 13: 43 Page 7

Introduction 7

1.4.3 Singular integral equations

A singular integral equation is defined as an integral with the infinite limits or when
the kernel of the integral becomes unbounded at a certain point in the interval. As
for examples,

∞
u(x) = f (x) + λ u(t)dt
−∞

x 1
f (x) = u(t)dt, 0 < α < 1 (1.14)
0 (x − t)α

are classified as the singular integral equations.

1.4.4 Integro-differential equations

In the early 1900, Vito Volterra studied the phenomenon of population growth, and
new types of equations have been developed and termed as the integro-differential
equations. In this type of equations, the unknown function u(x) appears as the
combination of the ordinary derivative and under the integral sign. In the electrical
engineering problem, the current I (t) flowing in a closed circuit can be obtained in
the form of the following integro-differential equation,

t
dI 1
L + RI + I (τ)dτ = f (t), I (0) = I0 (1.15)
dt C 0

where L is the inductance, R the resistance, C the capacitance, and f (t) the applied
voltage. Similar examples can be cited as follows:

x
u (x) = f (x) + λ (x − t)u(t)dt, u(0) = 0, u (0) = 1, (1.16)
0

1

u (x) = f (x) + λ (xt)u(t)dt, u(0) = 1. (1.17)
0

Equations (1.15) and (1.16) are of Volterra type integro-differential equations,

whereas equation (1.17) Fredholm type integro-differential equations. These
terminologies were concluded because of the presence of indefinite and definite
integrals.

1.5 Converting Volterra equation to ODE

In this section, we shall present the technique that converts Volterra integral equa-
tions of second kind to equivalent ordinary differential equations. This may be
MM-165 CH001.tex 26/4/2007 13: 43 Page 8

8 Integral Equations and their Applications

 b(x)
achieved by using the Leibnitz rule of differentiating the integral a(x) F(x, t)dt
with respect to x, we obtain

b(x)
b(x)
d ∂F(x, t) db(x)
F(x, t)dt = dt + F(x, b(x))
dx a(x) a(x) ∂x dx
da(x)
− F(x, a(x)), (1.18)
dx

where F(x, t) and ∂F ∂x (x, t) are continuous functions of x and t in the domain
α ≤ x ≤ β and t0 ≤ t ≤ t1 ; and the limits of integration a(x) and b(x) are defined
functions having continuous derivatives for α ≤ x ≤ β. For more information the
reader should consult the standard calculus book including Rahman (2000). A
simple illustration is presented below:

x
x  
d dx
sin(x − t)u(t)dt = cos(x − t)u(t)dt + (sin(x − x)u(x))
dx 0 0 dx
 
d0
− (sin(x − 0)u(0))
dx

x
= cos(x − t)u(t)dt.
0

1.6 Converting IVP to Volterra equations

We demonstrate in this section how an initial value problem (IVP) can be trans-
formed to an equivalent Volterra integral equation. Let us consider the integral
equation

t
y(t) = f (t)dt (1.19)
0
∞
The Laplace transform of f (t) is defined as L{f (t)} = 0 e−st f (t)dt = F(s). Using
this definition, equation (1.19) can be transformed to

1
L{y(t)} = L{f (t)}.
s
t t
In a similar manner, if y(t) = 0 0 f (t)dtdt, then

1
L{y(t)} = L{f (t)}.
s2
This can be inverted by using the convolution theorem to yield

t
y(t) = (t − τ)f (τ)dτ.
0
MM-165 CH001.tex 26/4/2007 13: 43 Page 9

Introduction 9

If

t
t
t
y(t) = ··· f (t)dtdt · · · dt
0 0
 0

n-fold integrals

then L{y(t)} = s1n L{f (t)}. Using the convolution theorem, we get the Laplace
inverse as

t
(t − τ)n−1
y(t) = f (τ)dτ.
0 (n − 1)!
Thus the n-fold integrals can be expressed as a single integral in the following
manner:

t
t
t
t
(t − τ)n−1
··· f (t)dtdt · · · dt = f (τ)dτ. (1.20)
(n − 1)!
0 0 0
 0

n-fold integrals

This is an essential and useful formula that has enormous applications in the
integral equation problems.

1.7 Converting BVP to Fredholm integral equations

In the last section we have demonstrated how an IVP can be transformed to an
equivalent Volterra integral equation. We present in this section how a boundary
value problem (BVP) can be converted to an equivalent Fredholm integral equation.
The method is similar to that discussed in the previous section with some exceptions
that are related to the boundary conditions. It is to be noted here that the method
of reducing a BVP to a Fredholm integral equation is complicated and rarely used.
We demonstrate this method with an illustration.

Example 1.1
Let us consider the following second-order ordinary differential with the given
boundary conditions.

y (x) + P(x)y (x) + Q(x)y(x) = f (x) (1.21)

with the boundary conditions

x=a: y(a) = α
y=b: y(b) = β (1.22)

where α and β are given constants. Let us make transformation

y (x) = u(x) (1.23)

MM-165 CH001.tex 26/4/2007 13: 43 Page 10

10 Integral Equations and their Applications

Integrating both sides of equation (1.23) from a to x yields

x
 
y (x) = y (a) + u(t)dt (1.24)
a

Note that y (a) is not prescribed yet. Integrating both sides of equation (1.24) with
respect to x from a to x and applying the given boundary condition at x = a, we find

x
x
y(x) = y(a) + (x − a)y (a) + u(t)dtdt
a a

x
x
= α + (x − a)y (a) + u(t)dtdt (1.25)
a a

and using the boundary condition at x = b yields

b
b
y(b) = β = α + (b − a)y (a) + u(t)dtdt,
a a

and the unknown constant y (a) is determined as

b
b
β−α 1
y (a) = − u(t)dtdt. (1.26)
b−a b−a a a

Hence the solution (1.25) can be rewritten as


β−α 1 b
b
y(x) = α + (x − a) − u(t)dtdt
b−a b−a a a

x
x
+ u(t)dtdt (1.27)
a a

Therefore, equation (1.21) can be written in terms of u(x) as


x
u(x) = f (x) − P(x) y (a) + u(t)dt
a

x
x

−Q(x) α + (x − a)y (a) + u(t)dtdt (1.28)
a a

where u(x) = y (x) and so y(x) can be determined, in principle, from equation (1.27).
This is a complicated procedure to determine the solution of a BVP by equivalent
Fredholm integral equation.
MM-165 CH001.tex 26/4/2007 13: 43 Page 11

Introduction 11

A special case
If a = 0 and b = 1, i.e. 0 ≤ x ≤ 1, then

x
x
y(x) = α + xy (0) + u(t)dtdt
0 0

x

= α + xy (a) + (x − t)u(t)dt
0

And hence the unknown constant y (0) can be determined as

1

y (0) = (β − α) − (1 − t)u(t)dt
0

x
1
= (β − α) − (1 − t)u(t)dt − (1 − t)u(t)dt
0 x

And thus we have


x

u(x) = f (x) − P(x) y (0) + u(t)dt
0

x
−Q(x) α + xy (0) + (x − t)u(t)dt
0

1
u(x) = f (x) − (β − α)(P(x) + xQ(x)) − αQ(x) + K(x, t)u(t)dt
0
(1.29)

where the kernel K(x, t) is given by

(P(x) + tQ(x))(1 − x) 0 ≤ t ≤ x
K(x, t) = (1.30)
(P(x) + xQ(x))(1 − t) x ≤ t ≤ 1

It can be easily verified that K(x, t) = K(t, x) confirming that the kernel is symmetric.
The Fredholm integral equation is given by (1.29).

Example 1.2
Let us consider the following boundary value problem.

y (x) = f (x, y(x)), 0≤x≤1

y(0) = y0 , y(1) = y1 (1.31)

Integrating equation (1.31) with respect to x from 0 to x two times yields

x
x
y(x) = y(0) + xy (0) + f (t, y(t))dtdt
0 0

x
= y0 + xy (0) + (x − t)f (t, y(t))dt (1.32)
0
MM-165 CH001.tex 26/4/2007 13: 43 Page 12

12 Integral Equations and their Applications

To determine the unknown constant y (0), we use the condition at x = 1, i.e.

y(1) = y1 . Hence equation (1.32) becomes

1

y(1) = y1 = y0 + y (0) + (1 − t)f (t, y(t))dt,
0

and the value of y (0) is obtained as

1
y (0) = (y1 − y0 ) − (1 − t)f (t, y(t))dt.
0

Thus, equation (1.32) can be written as

1
y(x) = y0 + x(y1 − y0 ) − K(x, t)f (t, y(t))dt, 0≤x≤1 (1.33)
0

in which the kernel is given by

t(1 − t) 0 ≤ t ≤ x
K(x, t) = (1.34)
x(1 − t) x ≤ t ≤ 1.

Once again we can reverse the process and deduce that the function y which satisfies
the integral equation also satisfies the BVP. If we now specialize equation (1.31)
to the simple linear BVP y (x) = −λy(x), 0 < x < 1 with the boundary conditions
y(0) = y0 , y(1) = y1 , then equation (1.33) reduces to the second kind Fredholm
integral equation

1
y(x) = F(x) + λ K(x, t)y(t)dt, 0≤x≤1
0

where F(x) = y0 + x(y1 − y0 ). It can be easily verified that K(x, t) = K(t, x) con-
firming that the kernel is symmetric.

Example 1.3
As the third example, consider the following boundary value problem

y (x) + y(x) = x, 0 < x < π/2

y(0) = 1, y(π/2) = π (1.35)

The analytical solution of the above problem is simply y(x) = cos x +

π
2 sin x + x. We want to reduce it into Fredholm integral equation.
MM-165 CH001.tex 26/4/2007 13: 43 Page 13

Introduction 13

Integrating the differential equation with respect to x from 0 to x twice and using
the boundary conditions, we obtain

x
x
 x3
y(x) = xy (0) + − y(t)dtdt
6 0 0

x
x3
= xy (0) + − (x − t)y(t)dt
6 0

Using the boundary condition at x = π2 , the unknown constant y (0) can be

obtained as


 π2 2 π/2
y (0) = 2 − + (π/2 − t)y(t)dt.
24 π 0
With this information the equation for y(x) can be put in the form of Fredholm
integral equation

π/2
y(x) = f (x) + K(x, t)y(t)dt,
0

π2 x3
where f (x) = 2x − 24 x + 6 and the kernel is given by

2t
π (π/2 − t) 0≤t≤x
K(x, t) = (1.36)
π (π/2 − x) x ≤ t ≤ π/2,
2x

which can be easily shown that the kernel is symmetric as before.

1.8 Types of solution techniques

There are a host of solution techniques that are available to solve the integral
equations. Two important traditional methods are the method of successive approxi-
mations and the method of successive substitutions. In addition, the series method
and the direct computational method are also suitable for some problems. The
recently developed methods, namely the Adomian decomposition method (ADM)
and the modified decomposition method, are gaining popularity among scientists
and engineers for solving highly nonlinear integral equations. Singular integral
equations encountered by Abel can easily be solved by using the Laplace transform
method. Volterra integral equations of convolution type can be solved using the
Laplace transform method. Finally, for nonlinear problems, numerical techniques
will be of extremely useful to solve the highly complicated problems.
This textbook will contain two chapters dealing with the integral equations
applied to classical problems and the modern advanced problems of physical
interest.
MM-165 CH001.tex 26/4/2007 13: 43 Page 14

14 Integral Equations and their Applications

1.9 Exercises

1. Classify each of the following integral equations as Volterra or Fredholm integral

equation, linear or nonlinear, and homogeneous or nonhomogeneous:

1
(a) u(x) = x + (x − t)2 u(t)dt
0

x
(b) u(x) = e +x
t 2 u2 (t)dt
0

π/2
(c) u(x) = cos x + cos xu(t)dt
0

1
x 1 1
(d) u(x) = 1 + dt
4 0 x + t u(t)
2. Classify each of the following integro-differential equations as Volterra integro-
differential equations or Fredholm integro-differential equations. Also deter-
mine whether the equation is linear or nonlinear.

x

(a) u (x) = 1 + e−2t u3 (t)dt, u(0) = 1
0

x
x2

(b) u (x) = − (x − t)u2 (t)dt, u(0) = 1, u (0) = 0
2 0

π/2

(c) u (x) = sin x − x + xtu (t)dt
0
u(0) = 1, u (0) = 0, u (0) = −1

3. Integrate both sides of each of the following differential equations once from 0
to x, and use the given initial conditions to convert to a corresponding integral
equations or integro-differential equations.

(a) u (x) = u2 (x), u(0) = 4

(b) u (x) = 4xu2 (x), u(0) = 2, u (0) = 1
(c) u (x) = 2xu(x), u(0) = 0, u (0) = 1.

4. Verify that the given function is a solution of the corresponding integral

equations or integro-differential equations:

x
(a) u(x) = x − (x − t)u(t)dt, u(x) = sin x
0

x
(b) (x − t)2 u(t)dt = x3 , u(x) = 3
0

x
(c) (x − t)u(t)dt = x3/2 , u(x) = 3/2.
0
MM-165 CH001.tex 26/4/2007 13: 43 Page 15

Introduction 15

5. Reduce each of the Volterra integral equations to an equivalent initial value

problem:

x
(a) u(x) = x − cos x + (x − t)u(t)dt
0

x
(b) u(x) = x4 + x2 + 2 (x − t)2 u(t)dt
0

x
1
(c) u(x) = x2 + (x − t)3 u(t)dt.
6 0

6. Derive an equivalent Volterra integral equation to each of the following initial

value problems:

(a) y + 5y + 6y = 0, y(0) = 1, y (0) = 1

(b) y + y = sin x, y(0) = 0, y (0) = 0
(c) y + 4y = x, y(0) = 0, y (0) = 0, y (0) = 1
d 4y d 2y
(d) + 2 = 2ex , y(0) = 2, y (0) = 2, y (0) = 1, y (0) = 1.
dx4 dx
7. Derive the equivalent Fredholm integral equation for each of the following
boundary value problems:

(a) y + 4y = sin x, 0 < x < 1, y(0) = 0, y(1) = 0


(b) y + 2xy = 1, 0 < x < 1, y(0) = 0, y(1) = 0

(c) y + y = x, 0 < x < 1, y(0) = 1, y(1) = 0

(d) y + y = x, 0 < x < 1, y(0) = 1, y (1) = 0.

References
[1] Lovitt, W.V., Linear Integral Equations, Dover: New York, 1950.
[2] Piaggio, H.T.H., An Elementary Treatise on Differential Equations and their
Applications, G. Bell and Sons, Ltd.: London, 1920.
[3] Rahman, M., Applied Differential Equations for Scientists and Engineers,
Vol. 1: Ordinary Differential Equations, WIT Press: Southampton, 1994.
[4] Tricomi, F.G., Integral Equations, Interscience: New York, 1957.
[5] Wazwaz, A.M., A First Course in Integral Equations, World Scientific:
Singapore, 1997.
[6] Wylie, C.R. & Barrett, L.C., Advanced Engineering Mathematics, McGraw-
Hill: New York, 1982.