Chapter 13

Volterra Integral Equations

13.1 Introduction
In mathematics, an integral equation is an equation in which an unknown
function appears under an integral sign. Many initial and boundary value
problems associated with ODE or PDE can be transformed into problems
of solving some approximate integral equations [4, page 1]. There is a close
connection between differential and integral equations, and some problems
may be formulated either way. In 1825, an Italian mathematician, Niels
Henrik Abel attacked a generalized version of the tautochrone problem (a
problem connected with determining a curve along which a heavy particle,
sliding without friction, descends to its lowest position). He derived the
solution of problem in term of Integral equation, is popularly known as Abel’s
integral equation which we will discuss later.
An integral equation can be classified as a linear or a nonlinear integral
equation as we have in ODEs and PDEs. The most used integral equations
are Volterra and Fredholm integral equations. The major integral equations
can be classified into four types, namely

• Volterra integral equation.

• Fredholm integral equation.

• Integro-differential equation.

• Singular integral equation.

The most basic type of integral equation is an equation of the form

Zb
f (x) = K(x, t) ϕ(t) dt.
a

160
Samir K. Bhowmik: VIE 161

Here ϕ(t) is an unknown function, f and K are known functions. K is often

called the kernel function of the integral. This integral equation is known as
a Fredholm equation. If the unknown function occurs both inside and outside
of the integral, it is known as a Fredholm equation of the second type:

Zb
ϕ(x) = f (x) + λ K(x, t) ϕ(t) dt.
a

where λ is an unknown factor, which plays the same role as the eigenvalue
in linear algebra.
There is another type of integral equation with variable limit (if one
limit of integration is a variable) and is known as Volterra equation. In this
Chapter, we shall be concerned with the nonhomogeneous Volterra integral
equation of the second kind of the form
Z t
x(t) = f (t) + λ K(t, s), x(s)ds, (13.1)
t0

where K(t, s) is the kernel function of the integral equation, f (t) a continuous
function of t, and λ a parameter. Here K(t, s) and f (t) are the given functions
but x(t) is an unknown function that needs to be determined. This is known
as Volterra integral equation of second kind. If the known function f (t) = 0,
it is called a homogeneous integral equation. If f (t) 6= 0, it is called an
inhomogeneous integral equation.
The nonhomogeneous Volterra integral equations of the first kind is de-
fined as Z t
f (t) = λ K(t, s)x(s)ds.
t0

A linear Volterra integral equation (LVIE) of first kind can be written as

Z t
f (t) = K(t, s)x(s)ds. (13.2)
t0

If df
dt
= f 0 (t), ∂K
∂t
= Kt (t, s) exists and are continuous, the LVIE of first kind
can be converted to the LVIE of second kind. Differentiating (13.2) with
respect to t, we have
Z t
K(t, t)x(t) + Kt (t, s)x(s)ds = f 0 (t),
t0

which can be written as

t
f 0 (t)
Z
Kt (t, s)
x(t) + x(s)ds = ,
t0 K(t, t) K(t, t)
Samir K. Bhowmik: VIE 162

if K(t, t) 6= 0. The above equation shows a relation between a LVIE of the

first and the second kind. In general, VIE can be written as
Z t
x(t) = f (t) + λ K(t, s, x(s))ds. (13.3)
t0

There are several number of solution techniques to deal with Volterra

integral equations. The VIE can be solved by using the Picard’s method of
successive approximations.

13.2 The Method of Successive Approxima-

tions
We replace the unknown function x(t) under integral sign of (13.3) by any se-
lective real-valued continuous function x0 (t), called the initial approximation
to get Z t
x1 (t) = f (t) + λ K(t, s, x0 (s))ds.
t0

Here x1 (t) is a continuous function if f , K and x0 are continuous. Similarly

second approximation can be made by
Z t
x2 (t) = f (t) + λ K(t, s, x1 (s))ds,
t0

and the third by

Z t
x3 (t) = f (t) + λ K(t, s, x2 (s))ds,
t0

and so on. We write the recurrence formula as

Z t
xn (t) = f (t) + λ K(t, s, xn−1 (s))ds,
t0

where n = 1, 2, · · · , and x0 is a known real valued function. The following

theorem shows that the above sequence converges to the solution of VIE
(13.3).
Theorem 13.1. Enunciate a set of conditions for which VIE of the second
kind has a unique solution and prove it.
Statement. For the VIE of the second kind
Z t
x(t) = f (t) + K(t, s, x(s))ds, t ∈ [t0 , t0 + a] (13.4)
t0

assume
Samir K. Bhowmik: VIE 163

1. f (t) is a real continuous function on [t0 , t0 + a], a > 0;

2. K(t, s, x(s)) is a continuous function on
∆ : t0 ≤ s ≤ t ≤ t0 + a, |x − f (t)| ≤ b, b > 0;

3. K(t, s, x(s)) satisfies the Lipschitz condition

|K(t, s, x(s)) − K(t, s, y(s))| ≤ L|x − y|, L > 0,
in ∆.
Then, there exists a unique solution x(t) of (13.4) for all t ∈ [t0 , t0 + δ],
δ ∈ min{a, b/M }, where M = sup |K(t, s, x)|, (t, s, x) ∈ ∆.
Proof. We shall apply the method of successive approximations, constructing
the sequence {xn (t)} as follows:
x0 (t) = f (t)
Z t
x1 (t) = f (t) + k(t, s, x0 (s))ds
t0
..
. (13.5)
Z t
xn (t) = f (t) + k(t, s, xn−1 (s))ds, n ≥ 1.
t0

By mathematical induction we can see that all xn (t) are continuous functions
on [t0 , t0 + δ], δ given in the statement.
To prove the uniform convergence of the sequence {xn (t)}, we shall con-
sider the associated series
X∞
[xn (t) − xn−1 (t)] . (13.6)
n=1

Now
Z t Z t
|x1 (t) − x0 (t)| = | K(t, s, x0 (s))ds| ≤ |K(t, s, x0 (s))|ds
t0 t0
Z t
≤ M ds = M (t − t0 ), t0 ∈ [t0 , t0 + δ]
t0

To obtain the estimates for |xn (t) − xn−1 (t)|, n ≥ 2. We note that (13.5)
gives
Z t Z t
|xn (t) − xn−1 (t)| = | K(t, s, xn−1 (s))ds − K(t, s, xn−2 (s))ds|
t0 t0
Z t
≤ |K(t, s, xn−1 (s)) − K(t, s, xn−2 (s))|ds
t0
Z t
≤L |xn−1 (s) − xn−2 (s)|ds t0 ∈ [t0 , t0 + δ].
t0
Samir K. Bhowmik: VIE 164

It is easy to see that

(t − t0 )2
|x2 (t) − x1 (t)| ≤ M L , t ∈ [t0 , t0 + δ],
2!
(t − t0 )3
|x3 (t) − x2 (t)| ≤ M L2 , t ∈ [t0 , t0 + δ].
3!
In general we write

(t − t0 )n
|xn (t) − xn−1 (t)| ≤ M Ln−1 , t ∈ [t0 , t0 + δ],
n!
which can be established through induction hypothesis.
Now
∞ ∞
X X (t − t0 )n M  L(t−t0 )
M Ln−1

|xn (t)−xn−1 (t)| ≤ = e −1 , t ∈ [t0 , t0 +δ],
n=1 n=1
n! L

which is finite. Consequently, the infinite series ∞

P
n=1 |xn (t) − xn−1 (t)| con-
verges uniformly on [t0 , t0 + δ].
Now we set n
X
xn (t) = x0 + |xk (t) − xk−1 (t)|,
k=1

which implies that the sequence {xn (t)} and the series x0 + ∞
P
|x (t) −
P∞ k
k=1
xk−1 (t)| have the same convergence properties. Since the series x0 + k=1 |xk (t)−
xk−1 (t)| is uniformly convergent on [t0 , t0 + δ], so is the sequence {xn (t)} over
the same interval. Hence the sequence {xn (t)} is uniformly convergent on
[t0 , t0 + δ]. We denote
x(t) = lim xn (t).
n→∞

The function x(t) is continuous in [t0 , t0 +δ] and it can be seen that it satisfies
(13.4). Indeed, considering n → ∞, (13.5) can be written as
Z t
x(t) = lim xn (t) = f (t) + lim K(t, s, xn−1 (s))ds.
n→∞ n→∞ t0

which follows (13.4) since

Z t Z t
lim K(t, s, xn−1 (s))ds = K(t, s, x(s))ds, t ∈ [t0 , t0 + δ].
n→∞ t0 t0

The last relation holds, because

|K(t, s, xn−1 (s)) − K(t, s, x(s))| ≤ L|xn−1 (s) − x(s)|, t0 ≤ s ≤ t ≤ t0 + δ.

thus the existence of solution is proved.

Samir K. Bhowmik: VIE 165

To prove the uniqueness, let y(t) be a continuous solution of (13.4) in

t0 ≤ t ≤ t0 + δ. Then y(t) satisfies (13.4). Then
Z t
|y(t) − xn (t)| ≤ |K(t, s, y(s)) − K(t, s, xn−1 (s))|ds, t ∈ [t0 , t0 + δ]
t0
Z t
≤L |y(s) − xn−1 (s)|ds, t ∈ [t0 , t0 + δ].
t0

But Z t Z t
|y(t) − f (t)| ≤ |K(t, s, y(s))|ds ≤ M ds = M (t − t0 ),
t0 t0

and a simple mathematical induction gives

(t − t0 )n+1
|y(t) − xn (t)| ≤ M Ln , t ∈ [t0 , t0 + δ].
(n + 1)!

Hence,
y(t) = lim xn (t) = x(t) ≡ y(t).
n→∞

Thus the solution x(t) of (13.4) is unique.

Example 13.1. Solve the VIE

Z x
u(x) = x + (t − x)u(t)dt.
0
Rx x3
Rx
Letu0 (x) = x. Then u1 (x) = x + 0 (t − x)tdt = x − 3!
. u2 (x) = x + 0
(t −
3 3 5
x) t − t3! dt = x − x3! + x5! . Thus

n
X x2i+1
un (x) = (−1)i
i=0
(2i + 1)!

which converges to sin x as n → ∞.

13.2.1 Linear Volterra integral equation of second kind

Theorem 13.2 (Method of successive Substitution). Consider linear integral
equation (VLIE) of the second kind
Z t
x(t) = f (t) + K(t, s)x(s)ds, t ∈ [t0 , t0 + a] (13.7)
t0

where

1. f (t) is a given scalar continuous function on [t0 , t0 + a]

Samir K. Bhowmik: VIE 166

2. K(t, s) is a given scalar continuous function on t0 ≤ s ≤ t ≤ t0 + a,

which is known as kernel of the integral equation (13.7).
Prove that the VLIE (13.7) has a unique solution and solve it.
Proof. Let us construct the following sequence

x0 (t) = f (t)
Z t
x1 (t) = f (t) + K(t, s)f (s)ds
t
Z 0t Z t
x2 (t) = f (t) + K(t, s)f (s)ds + K2 (t, s)f (s)ds
t0 t0
..
. (13.8)
Z t Z t Z t
xn (t) = f (t) + K(t, s)f (s)ds + K2 (t, s)f (s)ds + · · · + Kn (t, s)f (s)ds
t0 t0 t0

where Z t
Kn (t, s) = K(t, u)Kn−1 (u, s), n≥2 (13.9)
s
with K1 (t, s) = K(t, s). From the general formula for recurrence
Z t
xn (t) = f (t) + K(t, s)xn−1 (s)ds, n ≥ 1,
t0

and it can be written as

n
Z tX
xn (t) = f (t) + Kj (t, s)f (s)ds, n ≥ 1, t ∈ [t0 , t0 + a]. (13.10)
t0 j=1

Now we show that the series

n
X
Kj (t, s) (13.11)
j=1

converges uniformly on t0 ≤ s ≤ t ≤ t0 + a. Since K(t, s) is continuous

in t0 ≤ s ≤ t ≤ t0 + a, there exists A > 0 such that |K(t, s)| < A for all
t0 ≤ s ≤ t ≤ t0 + a. Now for all t0 ≤ s ≤ t ≤ t0 + a
Z t
|K2 (t, s)| ≤ |K(t, u)K1 (u, s)|du ≤ A2 (t − s),
s
t
(t − s)2
Z
|K3 (t, s)| ≤ |K(t, u)K2 (u, s)|du ≤ A3 ,
s 2!
and, in general, we write
Z t
(t − s)n−1
|Kn (t, s)| ≤ |K(t, u)Kn−1 (u, s)|du ≤ An ,
s (n − 1)!
Samir K. Bhowmik: VIE 167

which can be established by induction argument.

Now
n ∞ ∞
X X − s)n−1
n (t
X (t − s)n
|Kj (t, s)| ≤ A =A An = A exp(A(t − s)),
j=1 n=1
(n − 1)! n=0
n!

which is finite for all t0 ≤ s ≤ t ≤ t0 +a. This estimate shows that the infinite
series (13.11) is uniformly convergent on t0 ≤ s ≤ t ≤ t0 + a. Denoting by
r(s, t) the sum of the series (13.11), we get
n
X
r(t, s) = Kj (t, s),
j=1

which is continuous on t0 ≤ s ≤ t ≤ t0 + a. The function r(t, s) is called

the resolvent kernel for the kernel function K(t, s) of (13.7).
To prove the uniform convergence of the sequence {xn (t)} we shall con-
sider the associate series
∞
X
(xn (t) − xn−1 (t)).
n=1

Let M = sup |f (t)| for all t ∈ [t0 , t0 + a]. Then it is easy to see that

n (t − t0 )n
|xn (t) − xn−1 (t)| ≤ M A (Prove it!!!).
n!
Thus
∞ ∞
X X (t − t0 )n
|xn (t) − xn−1 (t)| ≤ M An = M (eA(t−t0 ) − 1),
n=1 n=1
n!

which is finite. So, by Weierstrass M-test, the series ∞

P
n=1 (xn (t) − xn−1 (t))
is uniformly convergent on [t0 , t0 + a], and is the Nth partial sum xN (t) =
PN
n=1 (xn (t) − xn−1 (t)). We write

lim xn (t) = x(t),

n→∞

which is a continuous function on [t0 , t0 + a]. Thus as n → ∞ in (13.10) we

get Z t
x(t) = f (t) + r(t, s)f (s)ds,
t0

which is the solution of (13.7).

Let y(t) be a continuous solution of VLIE on [t0 , t0 + a]:
Z t
y(t) = f (t) + K(t, s)y(s)ds.
t0
Samir K. Bhowmik: VIE 168

Then Z t
|y(t) − xn (t)| ≤ |K(t, s)||y(s) − xn−1(s) |ds.
t0
Similar to above discussion we write
(t − t0 )n+1
|y(t) − xn (t)| ≤ M An+1 .
(n + 1)!
n+1 n+1
Since the series ∞ n+1 (t−t0 )
converges on [t0 , t0 +a], M An+1 (t−t0)
P
n=1 M A (n+1)! (n+1)!
→
0 as n → ∞. Hence
lim |y(t) − xn (t)| = 0,
n→∞
which implies that
y(t) = lim xn (t) = x(t).
n→∞
That is, y(t) ≡ x(t) on [t0 , t0 + a], which proves the uniqueness of solutions
of the VLIE.
Claim 13.1. The integral equation of the resolvent kernel is
Z t
r(t, s) = K(t, s) + K(t, u)r(u, s)du.
s

Proof. By the definition of the resolvent kernel

∞
X
r(t, s) = Kj (t, s),
j=1

and
Z t
Kn (t, s) = K(t, u)Kn−1 (u, s)du, n ≥ 2, with K1 (t, s) = K(t, s).
s

Then,
∞
X ∞
X
r(t, s) = Kj (t, s) = K1 (t, s) + Kj (t, s)
j=1 j=2
∞ Z
X t
= K(t, s) + K(t, u)Kj−1 (u, s)du
j=2 s
∞
!
Z t X
= K(t, s) + K(t, u) Kj−1 (u, s) du
s j=2
∞
!
Z t X
= K(t, s) + K(t, u) Kj (u, s) du
s j=1
Z t
= K(t, s) + K(t, u)r(u, s)du,
s

completes the proof.

Samir K. Bhowmik: VIE 169

Example 13.2. Solve the integral equation

Z x
u(x) = 1 + u(t)dt
0

using successive substitution.

Here it is easy to see that by successive substitution a general solution can
be written as
Z x Z xZ x
2
u(x) = f (x) + λ K(x, t)f (t)dt + λ K(x, t)K(t, t1 )f (t1 )dtdt1 + · · · .
0 0 0

Here for this example λ = 1, f (x) = 1, K(x, t) = 1. Thus

Z x Z xZ x Z xZ xZ x
2
u(x) = 1 + dt + dt + dt3 + · · ·
0 0 0 0 0 0
x2 x3
=1+x+ + + · · · = ex .
2! 3!

13.3 Laplace transform method

Laplace transformation can be used to solve Volterra integral equation if the
integral is of convolution type. Here we consider
Z t
x(t) = f (t) + K(t − s)x(s)ds, t ∈ [t0 , t0 + a]. (13.12)
0

Applying the Laplace transform on (13.12) we have

L{x(t)} = L{f (t)} + L{K(t)}L{x(t)},

and thus the solution is

L{f (t)}
L{x(t)} = ,
1 − L{K(t)}

where Z ∞
L{x(t)}(s) = e−st x(t)dt.
0

Thus the solution of x(t) of (13.12) can be written as

Z t
x(t) = ψ(t − s)f (s)ds,
0

where  
−1 1
ψ(x) = L .
1 − L{K(t)}
Samir K. Bhowmik: VIE 170

Example 13.3. Solve the VIE

Z t
u(t) = f (t) + et−s u(s)ds
0

using Laplace transform and successive approximations.

 
1
L{u(t)} = L{f (t)} + L{K(t)}L{u(t)} = 1 + L{f (t)}.
s−2
Thus the inverse Laplace transform gives
Z t Z t
2(t−s)
e2(t−s) f (s)ds,


u(t) = δ(t − s) + e f (s)ds = f (t) +
0 0

where δ(t) is the Dirac delta function.

Example 13.4. Find the resolvent kernel and hence solve

Z x
ϕ(x) = x + (ξ − x)ϕ(ξ)dξ.
0

Here, f (x) = x and K(x, ξ) = K1 (x, ξ) = ξ − x. Thus

Z x Z x
(ξ − x)3
K2 (x, ξ) = K(x, u)K1 (u, ξ)du = (u − x)(u − ξ)du = (−1)2−1 ,
ξ ξ 3!
Z x Z x
(ξ − x)3 (ξ − x)5
K3 (x, ξ) = K(x, u)K2 (u, ξ)du = − (u−x) du = (−1)3−1 .
ξ ξ 3! 5!
So by mathematical induction we assume that
Z x
(ξ − x)2n−1
Kn (x, ξ) = K(x, u)Kn−1 (u, ξ)du = (−1)n−1 . (13.13)
ξ (2n − 1)!

Then
x x
(ξ − u)2n−1 (ξ − x)2n+1
Z Z
Kn+1 (x, ξ) = K(x, u)Kn (u, ξ)du = (u−x)(−1)n−1 du = (−1)n .
ξ ξ (2n − 1)! (2n + 1)!

Hence the relation (13.13) holds for all n ≥ 1. So the resolvent kernel for
the integral equation is
∞ ∞ 2n−1
n−1 (ξ − x)
X X
r(x, ξ) = Kn (x, ξ) = (−1) = sin(ξ − x)
n=1 n=1
(2n − 1)!

Hence the required solution is

Z x Z x
ϕ(x) = x + r(x, ξ)f (ξ)dξ = x + ξ sin(ξ − x)dξ = sin x.
0 0
Samir K. Bhowmik: VIE 171

13.4 The Adomian Decomposition Method

The Adomian decomposition method (ADM) is a non-numerical method for
solving nonlinear differential and integral equations, both ordinary and par-
tial. The general direction of this work is towards a unified theory for partial
differential equations (PDE). The method was developed by George Ado-
mian, chair of the Center for Applied Mathematics at the University of Geor-
gia. This method is a semi-analytical method. This is a power series method.
We express x(t) by
X∞
x(t) = xn (t)
n=0

with x0 (t) as the term outside integral sign. Here we consider the integral
equation (13.7). So x0 (t) = f (t). Substituting the series expression in
Z t
x(t) = f (t) + λ K(t, s)x(s)ds, t ∈ [t0 , t0 + a], (13.14)
t0

we get
∞
X Z t ∞
X
xn (t) = f (t) + λ K(t, s) xn (s)ds, t ∈ [t0 , t0 + a]. (13.15)
n=0 t0 n=0

The components x0 , x1 , x2 , · · · of the unknown function x(t) can be deter-

mined by
Z t Z t
x0 (t) = f (t), x1 (t) = λ K(t, s)x0 (s)ds, · · · , xn (t) = λ K(t, s)xn−1 (s)ds, · · · .
t0 t0

Thus a recurrence formula for the Adomian decomposition method can be

written as
Z t
x0 (t) = f (t), xn (t) = λ K(t, s)xn−1 (s)ds, n ≥ 1.
t0
Rt
Example 13.5. Solve the VIE x(t) = t + 0 (s − t)x(s)ds by using Adomian
decomposition method.
Considering
X∞
x(t) = xn (t),
n=0

and then substituting in the integral equation we have

∞
X Z t ∞
X
xn (t) = t + (s − t) xn (s)ds.
n=0 0 n=0
Samir K. Bhowmik: VIE 172

Thus
x0 (t) = t,
t
t3
Z
x1 (t) = (s − t)x0 (s)ds = − ,
0 3!
Z t 5
t
x2 (t) = (s − t)x1 (s)ds = ,
0 5!
..
.
Thus the solution of the VIE is
∞
X t3 t5
x(t) = xi (t) = t − + + · · · = sin t.
i=0
3! 5!

13.5 Series Solution Method

Let u(t) be an analytic function, and has Taylor’s expansion around t = 0
by
X∞
u(t) = an tn
n=0
where an are constants and required to be determined. Substituting the
series expression of u(t) in (13.14) we have
X∞ Z t ∞
X
n
an t = f (t) + λ K(x, t) an sn ds.
n=0 0 n=0

It is easy to notice that the above identity is a collection of expressions in

terms of ti , i = 0, 1, 2, 3, · · · . One can compare the coefficients of ti , to
evaluate ai values.
Example 13.6. Consider the Vloterra Integral equation
Z t
u(t) = 1 + 2 sin t − u(s)ds.
0
P∞ n
Substituting u(t) = n=0 an t we get
∞ ∞ ∞
X X t2n+1 X tn+1
an tn = 1 + 2 − an .
n=0 n=0
(2n + 1)! n=0 (n + 1)!
Thus comparing the coefficients we have a0 = 1, a1 = 2 − a0 , a2 = − a21 ,
a3 = − 13 − a32 , and so on. Thus we get the series solution
t2 t3 t4 t5
u(t) = 1 + t − − + +
2! 3! 4! 5!
= sin t + cos t.
Now applying Laplace transformation one may easily verify the result.
Samir K. Bhowmik: VIE 173

13.6 VIE and LDE

Now consider a n−th order differential equation of the form

dn y dn−1 y dn−2 y
+ a 1 (x) + a 2 (x) + · · · + an (x)y = f (x)
dxn dxn−1 dxn−2
with y i (0) = ci , where i = 0, 1, · · · n − 1 and ai (x) are continuous functions.
Consider
dn y
= u(x).
dxn
Integrating w.r.t. x Z x
dn−1 y
= u(x)dx + cn−1 .
dxn−1 0
Thus, integrating n − 1 times
Z xZ x Z x
xn−1 xn−2
y(x) = ··· u(x)dxdx · · · dx+cn−1 +cn−2 n+· · · c1 x+c0 .
0 0 0 (n − 1)! (n − 2)!

Substituting all these values to the differential equation we have

Z x
(x − t)2 (x − t)n−1

u(x) + a1 (x) + a2 (x)(x − t) + a3 (x) + · · · + +an (x) u(x)dx
0 2! (n − 1)!
xn−1
= F (x) − cn−1 a1 (x) − (cn−1 + cn−2 )a2 (x) − · · · − (cn−1 + · · · + c1 x + c0 )an (x)
(n − 1)!

which is a VIE of second kind.

Example 13.7. Reduce the IVP

y 00 (x) + 4y(x) = sin x; at x = 0, y(0) = 0, y 0 (0) = 0,

to VIE of the second kind and then find its solution.

Integrating twice we get
Z xZ x Z x
y(x) = u(t)dtdt = (x − t)u(t)dt.
0 0 0

Thus the ODE becomes

Z x
u(x) + 4 (x − t)u(t)dt = sin x.
0

Applying Laplace transform one may easily solve the integral equation and
the solution is  
1 1
y(x) = sin x − sin 2x .
3 2
Samir K. Bhowmik: VIE 174

13.7 Exercises
Exercise 13.1. Find the resolvent kernel and hence solve the VIE
Rx
1. ϕ(x) = 1 + 0 (ξ − x)ϕ(ξ)dξ.
Here f (x) = 1, K(x, ξ) = ξ − x. We get the resolvent kernel r(x, ξ) =
sin(ξ − x). Thus the solution is
ϕ(x) = cos x.
Rx
2. ϕ(x) = 1 + 0 ϕ(ξ)dξ.
Here f (x) = 1, K(x, ξ) = 1. We get the resolvent kernel r(x, ξ) =
exp(x − ξ). Thus the solution is
ϕ(x) = exp(x).
Rx
3. ϕ(x) = sin x + 2 0 ex−ξ ϕ(ξ)dξ.
Here f (x) = sin x, K(x, ξ) = 2ex−ξ . We get the recursive kernel
n−1
Kn (x, ξ) = 2n exp(x − ξ) (x−ξ)
(n−1)!
for all n ∈ N, and thus the resolvent
3(x−ξ)
kernel is r(x, ξ) = 2e , and the solution is
1
2 sin x − cos x + e3x .


ϕ(x) =
5
Rx
4. ϕ(x) = ex sin x + 0 2+cos x
2+cos ξ
ϕ(ξ)dξ.
Here f (x) = e sin x, K(x, ξ) = 2+cos
x x
2+cos ξ
and λ = 1. We get the recursive
n−1
x (x−ξ)
kernel Kn (x, ξ) = 2+cos
2+cos ξ (n−1)!
for all n ∈ N, and thus the resolvent
2+cos x x−ξ
kernel is R(x, ξ; λ) = 2+cos ξ e , λ = 1, and the solution is
 
x x 3
ϕ(x) = e sin x + e (2 + cos x) ln .
2 + cos x
Rx 1+x2
5. ϕ(x) = 1 + x2 + 0 1+ξ 2
ϕ(ξ)dξ.
2
Here f (x) = 1 + x2 , K(x, ξ) = 1+x 1+ξ 2
and λ = 1. We get the recursive
1+x2 (x−ξ)n−1
kernel Kn (x, ξ) = 1+ξ2 (n−1)! for all n ∈ N, and thus the resolvent
2
kernel is R(x, ξ; λ) = 1+x
1+ξ 2
ex−ξ , λ = 1, and the solution is

ϕ(x) = (1 + x2 )ex .
Rx
6. ϕ(x) = f (x) + λ 0 ex−s ϕ(ξ)dξ.
n−1
Here K(x, ξ) = ex−s . We get the recursive kernel Kn (x, ξ) = ex−s (x−s)
(n−1)!
for all n ∈ N, and thus the resolvent kernel is R(x, ξ; λ) = e(1+λ)(x−s) ,
and the solution is
Z x
ϕ(x) = f (x) + λ e(1+λ)(x−s) f (s)ds.
0
 Rt
7. x(t) = e−t + 0
e−(t−s) x(s)ds. The solution is x(t) = 1.
Rt
8. x(t) = cos t − t − 2 + 0 (s − t)x(s)ds. The solution is x(t) = − cos t −
sin t + 12 t sin t.
2 Rt 2 2 2
9. x(t) = et + 0 et −s x(s)ds. The solution is x(t) = et +t .

10. Reduce the IVP

y 00 (x) + 4y(x) = sin x, y(0) = 0, y 0 (0) = 0,

to Volterra integral equation of the second kind and then find it’s solu-
tion. (The solution is y(x) = 13 sin x − 61 sin 2x.)
Rx
11. Solve the LIE u(x) = 23 ex − 21 xex − 12 + 21 0 tu(t)dt.