Chapter 5

Solution to Nonlinear Integral

Equations

A nonlinear integral equation is an integral equation in which the unknown function appears
in a nonlinear manner. The nonlinearity may occur inside or outside of the integrand on in
both of the locations. It leads to an astonishing variety of new phenomenon related to the
characteristics of the solutions and to the method of solution.
Let Q(a, b) (resp. T (a, b)) be the square (resp. triangle) in the st-plane defined as

Q(a, b) = {(s, t) : [a, b] × [a, b]} (resp. T (a, b) = {(s, t) : a ≤ s ≤ b, a ≤ t ≤ s}) .

The most general type of nonlinear Fredholm integral equation is of the form

Zb
g(s) + λ F (s, t, f (t))dt = f (s),
a

in which we always assume the integrand F (s, t, f (t)) is continuous either on the set Q(a, b)×
R or on the smaller set Q(a, b) × [−M, M ] for a suitably large value of M .
Similarly, the most general type of nonlinear Volterra integral equation is of the form
Zs
g(s) + λ F (s, t, f (t))dt = f (s),
a

in which we always assume the integrand F (s, t, f (t)) is continuous either on the set T (a, b)×R
or on the smaller set T (a, b) × [−M, M ] for a suitably large value of M .
Note that, in these equations, the known function g(s) is always assumed to be continuous
on [a, b]. Moreover, the integrand F (s, t, f (t)) can factored as

F (s, t, f (t)) = K(s, t)G (t, f (t)) ,

where the kernel K(s, t) is either degenerate or symmetric & positive. In order to handle
these equations, we will have to use either advanced methods of functional analysis (e.g.

1
Banach’s fixed point theorem etc) and/or numerical techniques. Next, we present the tools
from functional analysis which we need to study the existence and uniqueness of solutions of
nonlinear Fredholm/Volterra integral equations.

5.1 Tools of the trade

• Metric, metric spaces and completeness: A metric on a set X is a mapping d : X × X → R
with the following properties:

(M1 ) d(x, y) ≥ 0 such that d(x, y) = 0 ⇐⇒ x = y, for all x, y ∈ X.

(M2 ) d(x, y) = d(y, x) for all x, y ∈ X.

(M3 ) d(x, z) ≤ d(x, y) + d(y, z) for all x, y, z ∈ X.

A metric is also called a distance function and d(x, y) is called the distance between x and y.
A metric space (X, d) is a set X endowed with the metric d. The elements of X are called
the points of X.
A sequence {xn } in a metric space (X, d) is called a Cauchy sequence in X if, for every
 > 0, there exists an integer N = N () such that

∀m, n ≥ N =⇒ d(xm , xn ) < .

The sequence {xn } is said to converge to x ∈ X if, for every  > 0, there exists an integer
N = N () such that
∀n ≥ N =⇒ d(xn , x) < .
A metric space X is said to be complete if every Cauchy sequence in X converges to a point
of X.
In particular, we note that the space C[a, b] of functions that are continuous on the closed
interval [a, b] is complete with respect to the metric

d (x(t), y(t)) = sup |x(t) − y(t)|.

t∈[a,b]

• Contraction mappings: Let (X, d) be a metric space. A map f : X → X is said to be a

contraction map or simply a contraction if there is a real number α, 0 < α < 1 such that

d(f (x), f (y)) ≤ αd(x, y)

for all x, y ∈ X. Note that the constant α does not in any way depend on the elements of X.
If X is a space of functions, then f is called contraction operator.

• Continuous mappings: Let (X, d) be a metric space. A map f : X → X is said to be

continuous at x0 ∈ X if, for every  > 0, there exists a δ > 0 such that

d(x, x0 ) < δ =⇒ d (f (x), f (x0 )) < .

2
Also, f is said to be continuous on Y ⊆ X, if it is continuous at every point of Y .

• Fixed points and fixed point theorems: Let f : X → X be a map, where X is metric space.
A point x0 ∈ X is said to be a fixed point of f if f (x0 ) = x0 .
Next, we prove the well-known Banach’s fixed point theorem.

Theorem 5.1.1 (Banach’s Fixed Point Theorem). Let (X, d) be a complete metric space
and f : X → X be a contraction mapping. Then, f has a unique fixed point x ∈ X such that

f (x) = x.

Proof: Since f is a contraction mapping on X, there is a real number α, 0 < α < 1 such that

d(f (x), f (y)) ≤ αd(x, y),

for all x, y ∈ X. So, using the fact f 2 (x) = f (f (x)), we have

d(f 2 (x), f 2 (y)) ≤ αd (f (x), f (y))

≤ α2 d(x, y)

Continuing this way, we find

d(f n (x), f n (y)) ≤ αn d(x, y), for all x, y ∈ X (1)

Now take any point x0 ∈ X and define a sequence {xn } in X by:

x1 = f (x0 ), x2 = f (x1 ) = f 2 (x0 ), . . . , xn = f (xn−1 ) = f n (x0 ), . . .

We show that {xn } is a Cauchy sequence in X. For this, we observe that, for any
m, n ∈ N, such that m > n, we have

d(xn , xm ) ≤ d(xn , xn+1 ) + d(xn+1 , dn+2 ) + . . . + d(xm−1 , xm )

≤ d(f n (x0 ), f n+1 (x0 )) + d(f n+1 (x0 ), f n+2 (x0 )) + . . . + d(f m−1 (x0 ), f m (x0 ))
≤ d(f n (x0 ), f n (x1 )) + d(f n+1 (x0 ), f n+1 (x1 ))+
. . . + d(f m−1 (x0 ), f m−1 (x1 )), ∵ f (x0 ) = x1 (2)

Using (1), we have from (2),

d(xn , xm ) ≤ αn d(x0 , x1 ) + αn+1 d(x0 , x1 ) + . . . + αm−1 d(x0 , x1 )

≤ αn + αn+1 + . . . + αm−1 d(x0 , x1 )

αn (1 − αm−n )
≤ d(x0 , x1 )
1−α
αn
≤ d(x0 , x1 ) → 0, as m, n → ∞, ∵ lim an = 0, if |a| < 1
1−α n→∞

3
 because α < 1. Hence {xn } is a Cauchy sequence. Since X is complete, {xn } converges
to a point of x ∈ X. Since every contraction mapping in continuous, we have

f (x) = f ( lim xn ) = lim f (xn ) = lim xn+1 = x

n→∞ n→∞ n→∞

Hence, f has a fixed point namely x.

To prove uniqueness of x, suppose that x0 is another fixed point of f . Then, x0 = f (x0 )
so that

d(x, x0 ) = d f (x), f (x0 )

≤ αd(x, x0 )
< d(x, x0 ), ∵ 0 < α < 1.

a contradiction because d(x, x0 ) = d(x, x0 ) . Hence, d(x, x0 ) = 0 implies x = x0 . Hence,

x is unique. 

Note that every contraction mapping is continuous. However, the converse may not be
true. However, it may happen that some power f n of f might be a contraction, even if f is
not. In this case, another fixed point theorem enunciated.

Theorem 5.1.2. Let (X, d) be a complete metric space and f : X → X be a map. If

f n : X → X, for some n ∈ N, is a contraction. Then, f has a unique fixed point.

Proof: For f : X → X, suppose that f n : X → X is a contraction. Then, there is a real

number α, 0 < α < 1 such that

d(f n (x), f n (y)) ≤ αd(x, y),

for all x, y ∈ X. So, by Theorem 5.1.1, f n has a fixed point x, say, so that

f n (x) = x

Now

d(f (x), x) = d(f (f n (x)), f n (x)) = d(f n (f (x)), f n (x))

≤ αd(f (x), x)
< d(f (x), x)

for 0 < α < 1. Hence, d(f (x), x) = 0 so that f (x) = x. Thus, f has fixed point.
For uniqueness of x, suppose that x0 is another fixed point of f . Then, x0 is a fixed
point of f n also. So,

d(x, x0 ) = d(f n (x), f n (x0 )) < αd(x, x0 ) < d(x, x0 ), for 0 < α < 1.

Hence, d(x, x0 ) = 0 implying x = x0 . This proves the uniqueness of x. 

4
• Lipschitz conditions: A real-valued function satisfies a Lipschitz condition on a set E if the
inequality
|f (x) − f (y)| ≤ C|x − y|
holds for all x, y ∈ E, where C is a constant that is independent of x & y. For example,
a function that has a bounded first derivative on E satisfies a Lipscitz condition as a con-
sequence of the Mean Value Theorem, which states that there exists a value x < c < y for
which
f (x) − f (y) = f 0 (c)(x − y).
A function of several variables can satisfy a Lipschitz condition on a set with respect
to one of the variable. In this chapter, F (s, t, z) of three variables that satisfy a Lipscitz
condition of the form
|F (s, t, z1 ) − F (s, t, z2 )| ≤ C|z1 − z2 |
where F (s, t, z) is defined on Q(a, b) × R. The constant C is independent of s, t, z1 and z2 .

5.2 Existence and uniqueness theorems

The fixed point theorems can be invoked to establish existence and uniqueness theorems for
nonlinear Fredholm/Volterra integral equations.

5.2.1 Nonlinear Fredholm integral equations

The following theorem employs Banach’s fixed point theorem to put forward the criteria for
existence and uniqueness of solution to a nonlinear Fredholm integral equation.

Theorem 5.2.1. Suppose that F (s, t, z) is defined and continuous on the set Q(a, b) × R and
that it satisfies a Lipschitz condition of the form

|F (s, t, z1 ) − F (s, t, z2 )| ≤ C|z1 − z2 |.

Suppose further that f ∈ C[a, b]. Then, the nonlinear Fredholm integral equation
Zb
g(s) + λ F (s, t, f (t))dt = f (s) (3)
a

1
has a unique solution on the interval [a, b] whenever |λ| < C(b−a) .

Proof: The space C[a, b] consisting of the real-values continuous function defined on [a, b] is
complete metric space with the metric

d(u, v) = sup |u(s) − v(s)|.

s∈[a,b]

If it can be shown that the operator

F : C[a, b] → C[a, b]

5
 defined by
Zb
Fu = g(s) + λ F (s, t, u(t))dt
a

is a contraction operator for the restricted values of λ prescribed in the statement

of the theorem, then it will follow immediately as an application of Banach’s fixed
point theorem that F has a unique fixed point i.e., that the integral equation (3) has
a unique solution. Observe that for any u, v ∈ C[a, b] and any s ∈ [a, b], we have

Zb
|Fu(s) − Fv(s)| = |λ| (F (s, t, u(t)) − F (s, t, v(t))) dt
a
Zb
≤ |λ| |F (s, t, u(t)) − F (s, t, v(t))|dt
a
Zb
≤ |λ|C sup |u(t) − v(t)| dt
t∈[a,b]
a
≤ |λ|C(b − a)d(u, v).

It follows directly from this inequality that

d(Fu, Fv) ≤ |λ|C(b − a)d(u, v).

Hence, if α = |λ|C(b − a) < 1, then F is a contraction operator from C[a, b] to C[a, b],
and the proof is complete. 

5.2.2 Nonlinear Volterra integral equations

Next, we study the existence and uniqueness of solution of nonlinear Volterra integral equa-
tions.

Theorem 5.2.2. Suppose that F (s, t, z) is defined and continuous on the set T (a, b) × R and
that it satisfies a Lipschitz condition of the form

|F (s, t, z1 ) − F (s, t, z2 )| ≤ C|z1 − z2 |.

Suppose further that f ∈ C[a, b]. Then, the nonlinear Volterra integral equation
Zs
g(s) + λ F (s, t, f (t))dt = f (s) (4)
a

has a unique solution on the interval [a, b] for every value of λ, where a ≤ s ≤ b.

6
Proof: The space C[a, b] consisting of the real-values continuous function defined on [a, b] is
complete metric space with the metric

d(u, v) = sup |u(s) − v(s)|.

s∈[a,b]

If it can be shown that the operator

V : C[a, b] → C[a, b]

defined by
Zs
Vu = g(s) + λ F (s, t, u(t))dt
a
is a contraction operator for any value of λ, then it will follow immediately as an
application of Banach’s fixed point theorem that V has a unique fixed point i.e., that
the integral equation (4) has a unique solution. Observe that for any u, v ∈ C[a, b]
and any s ∈ [a, b], we have
Zs
|Vu(s) − Vv(s)| = |λ| (F (s, t, u(t)) − F (s, t, v(t))) dt
a
Zs
≤ |λ| |F (s, t, u(t)) − F (s, t, v(t))|dt
a
Zs
≤ |λ|C|u(t) − v(t)| dt
a
≤ |λ|C(s − a)d(u, v).

After again invoking the Lipschitz condition and then integrating the result, we obtain
Zs
2 2
|V u(s) − V v(s)| = |λ| (F (s, t, Vu(t)) − F (s, t, Vv(t))) dt
a
Zs
≤ |λ|C |Vu(t) − Vv(t)|dt
a
|λ|2 C 2 (s − a)2
≤ d(u, v).
2!
In general, after n such integration, we obtain
|λ|n C n (s − a)n
|Vn u(s) − Vn v(s)| ≤ d(u, v),
n!
from which it follows that
|λ|n C n
sup |Vn u(s) − Vn v(s)| ≤ d(u, v) sup (s − a)n
s∈[a,b] n! s∈[a,b]

7
 |λ|n C n (b − a)n kn kn
d(Vn u, Vn v) ≤ d(u, v). ∵0< < 1 as lim =0
n! n! n→∞ n!

Hence, if n is sufficiently large, then Vn is a contraction operator from C[a, b] to itself,

and the proof is complete. 

5.3 Solution to nonlinear integral equations with degenerate

kernel
While linear integral equation with degenerate kernel is equivalent to a system of linear
equations, nonlinear integral equation with degenerate kernel is almost always equivalent to
a nonlinear system of algebraic equations.
We illustrate the method of solving nonlinear Fredholm integral equation with degenerate
kernel.

Example 5.3.1. Solve the following homogeneous integral equation:

Zπ
sin(s + t)f 2 (t)dt = f (s)
0

Solution. The given integral equation is

Zπ
(sin s cos t + cos s sin t)f 2 (t)dt = f (s) (5)
0

This can be written as

Zπ Zπ
2
sin s cos tf (t)dt + cos s sin tf 2 (t)dt = f (s) (6)
0 0

Set
Zπ
c1 = cos tf 2 (t)dt (7)
0

and
Zπ
c2 = sin tf 2 (t)dt (8)
0

Substituting the values of c1 and c2 from (7) & (8) respectively in (6) yields

c1 sin s + c2 cos s = f (s) (9)

8
Putting this value of f (t) = c1 sin t + c2 cos t in (7), we obtain
Zπ
c1 = cos t(c21 sin2 t + c22 cos2 t + 2c1 c2 sin t cos t)dt
0
Zπ Zπ Zπ
= c21 2
sin t cos tdt + c22 2
cos t(1 − sin )dt + 2c2 c2 cos2 t sin tdt
0 0 0
3 π 3 π 3 π
sin t sin t cos t 4
= c21 + c22 |sin t|π0 − c22 − 2c2 c2 = − c1 c2
3 0 3 0 3 0 3

This gives  
4 4
c1 1 − c2 = 0, c1 = c1 c2 . (10)
3 3
Similarly, putting f (t) = c1 sin t + c2 cos t in (8), we obtain
Zπ
c2 = sin t(c21 sin2 t + c22 cos2 t + 2c1 c2 sin t cos t)dt
0
Zπ Zπ Zπ
= c21 sin t(1 − cos t)dt + c22 2
cos t sin tdt + 2c2 c2 sin2 t cos tdt
0 0 0
π π 3 π
cos3 t cos3 t sin t 2 2
= c21 |sin t|π0 − c21 − c22 + 2c2 c2 = 2c21 − c21 + c22
3 0 3 0 3 0 3 3

This gives  
4 2 2
c + c2 c2 − 1 = 0 (11)
3 1 3
Case 1: Let c1 = 0.
Subcase 1.1: If c2 = 0, then we obtain (c1 , c2 ) = (0, 0).
Subcase 1.2: If c2 6= 0, then (11) implies that
 
2
c2 c2 − 1 = 0
3
2
c2 − 1 = 0, ∵ c2 6= 0
3
3
c2 = .
2
This gives us (c1 , c2 ) = 0, 23 .

9
 Case 2: Let c1 6= 0. Then, (10) implies that c2 = 34 . Plugging it in (11), we obtain

4 2 3
c − =0
3 1 8
4 2 3
c =
3 1 8
9
c21 =
32
3
c1 = ± √ .
4 2
Combining this with the previous two solutions, we obtain the following solutions:
     
3 3 3 3 3
(c1 , c2 ) = (0, 0), 0, , √ , , − √ ,
2 4 2 4 4 2 4
Thus, integral equation (5) has the following four possible solutions:
3 3 3
f (s) = 0, f (s) = cos s, f (s) = ± √ sin s + cos s.
2 4 2 4

10