International Journal of

Mathematical Modelling & Computations

Vol. 05, No. 04, Fall 2015, 361- 372

Approximation Solution of Two-Dimensional Linear Stochastic

Fredholm Integral Equation by Applying the Haar Wavelet
M. Fallahpour1 , K. Maleknejad2 and M. Khodabin3
1,2,3

Department of Mathematics, Karaj Branch, Islamic Azad University, Karaj, Iran.

Abstract.In this paper, we introduce an efficient method based on Haar wavelet to approximate a solution for the two-dimensional linear stochastic Fredholm integral equation. We also
give an example to demonstrate the accuracy of the method.

Received: 14 June 2015, Revised: 5 September 2013, Accepted: 18 November 2015.

Keywords: Aar wavelet, Two-dimensional Stochastic Fredholm Integral Equation,
Brownian Motion Process, Ito Integral.

Index to information contained in this paper

1
2
3
4
5

1.

Introduction
Haar Wavelets
Numerical Method
Numerical Example
Conclusion

Introduction

As we know, two dimensional ordinary integral equations provide an important tool

for modeling a numerous problems in engineering and science [6, 16]. The second
kind of two-dimensional integral equations appear in nonhomogeneous elasticity
and electrostatics, the Dorboux problem, contact problems for bodies with complex properties, radio wave propagation, the elastic problem of axial translation
of a rigid elliptical disc-inclusion, various physical and mechanical and biological
problems [1, 11, 4, 10, 20, 23, 25, 28]. Some numerical schemes have been inspected
for resolvent of two-dimensional ordinary integral equations by several probers.
Computational complexity of mathematical operations is the most important obstacle for solving ordinary integral equations in higher dimensionas.

1 [email protected]
2 [email protected]
3 Corresponding

author: [email protected]

c 2015 IAUCTB

http://www.ijm2c.ir

362

M. Fallahpour et al./ IJM2 C, 05 - 04 (2015) 361-372.

These include the Nystrom method, collocation method, Gauss product quadrature rule method, Galerkin method, using triangular fuctions, Legender polynomial
method, differential transform method, meshless method, Bernstein polynomials
method and Haar wavelet method [5, 7, 8, 9, 12, 13, 14, 15, 19, 21, 22, 24, 29, 30]. This
paper is first focused on proposing a generic framework for numerical solution of
two-dimensional ordinary linear Fredholm integral equations of second kind. The
use of the Haar wavelet for the numerical solution of linear integral equations
has previously been discussed in [3] and references therein. In [2] a new numerical method based on Haar wavelet is introduced for solution of nonlinear onedimensional Fredholm and Volterra integral equations. In [27] the Haar wavelet
method [2] is extended to numerical solution of integro-differential equation. In
[26] the Haar wavelet method [2, 27] is improved in terms of efficiency by introducing one-dimentional Haar wavelet approximation of the kernel function. The
method [3] is fundamentally different from the other numerical methods based on
Haar wavelet for the numerical solution of integral equations as it approximates
kernel function using Haar wavelet. Its easy to show that the Fredholm integral
form of the general hyperbolic differential equation [10] is given by two-dimentional
integral equation

1 1

g(x, y) = f (x, y) +

K1 (x, y, s, t)g(s, t)dsdt.

0

If we import statistical noise in the general hyperbolic differential equation [10],

we can obtain two-dimensional linear stochastic Fredholm integral equation of the
second kind, i.e.
1 1

K1 (x, y, s, t)g(s, t)dsdt

g(x, y) = f (x, y) +
0

(1)

1 1

K2 (x, y, s, t)g(s, t)dB(s)dB(t)

+
0

(x, y) [0, 1] [0, 1]

where the kernels K1 (x, y, s, t) and K2 (x, y, s, t) in (1.1) are known functions and
f (x, y) is also a known function whereas g(x, y) is unknown function and is called
the solution of two-dimensional stochastic integral equation.
Lemma 1.Put (t, s) = K(x, y, s, t)g(s, t). Let be a function in L2 ([0, 1]2 ).
Then there exists a sequence n of off-diagonal step functions such that [18]
b
lim

n a

| (t, s) n (t, s) |2 dtds = 0.

Definition 1. Let L2 ([0, 1]2 ). Then the double Wiener-Ito integral of is

M. Fallahpour et al./ IJM2 C, 05 - 04 (2015) 361-372.

363

defined as [18]
b

(t, s)dB(t)dB(s) = lim

a

n a

n (t, s)dB(t)dB(s)

in

L2 ().

Theorem 1. Let (t, s) L2 ([a, b]2 ). Then [18]

b

b [

(t, s)dB(t)dB(s) = 2
a

(t, s)dB(s) dB(t),

(2)

where is the symmetrization of that is defined by

s) = 1 ((t, s) + (s, t)) .
(t,
2
Also for the integral

b
a

B(t)dB(t) we have [18]

B(t)dB(t) =
a

2.

(3)

}
1{
B(b)2 B(a)2 (b a) .
2

(4)

Haar Wavelets

A wavelet family (j ,i (y))jN,iZ is an orthonormal subfamily of the Hilbert space

L2 (R) with the property that all function in the wavelet family are generated from
a fixed function called mother wavelet through dilations and translations.The
wavelet family satisfies the following relation
(
)
j ,i (y) = 2j/2 2j y i .
For Haar wavelet family on the interval [0, 1) we have:
{
h1 (y) =

1, f or y [0, 1)
0, otherwise,

(5)

and

1,
hi (y) = 1,

0,

f ory [, )
f ory [, )
otherwise
i = 2, 3, ...,

where
n =

n
,
m

m = 2 ,

n =

(n + 0.5)
,
m

= 0, 1, ...,

n =

(n + 1)
;
m

n = 0, 1, ..., m 1.

(6)

M. Fallahpour et al./ IJM2 C, 05 - 04 (2015) 361-372.

364

The integer indicats the level of the wavelet and n is the translation parameter.
Any square integrable function f (y) defined on [0, 1) can be expressed as follows:
f (y) =

ai hi (y),

i=1

where ai are real constants.

For approximation aim we consider a maximum value L of the integer , level of
the Haar wavelet in the above definition. The integer L is then called maximum
level of resolution. We also define integer M = 2L . Hence for any square integrable
function f (y) we have a finite sum of Haar wavelets as follows:
f (y)

2M

ai hi (y).

i=1

3.

Numerical Method

In this section proposed numerical method [3] will be discussed for two-dimensional
linear stochastic Fredholm integral equation of the second kind. In the first subsection, we state some results for efficient evalution of two-dimensional Haar wavelet
approximations. In the second subsection, we apply these results for finding numerical solutions equation (1.1).
For Haar wavelet approximation of a function f (x, y) of two real variables x and
y, we assume that the domain 0 x, y 1 is divided into a grid of size 2M 2N
using the following collocation points
xm =

yn =

3.1

m 0.5
, m = 1, 2, ..., 2M,
2M

(7)

n 0.5
, n = 1, 2, ..., 2N.
2N

(8)

Two-dimensional Haar wavelet system

A real-valued function G(x, y) of two real variables x and y can be approximated

using two-dimensional Haar wavelets basis as [3, 17]:
G(x, y)

2M
2N

bp,q hp (x)hq (y).

(9)

p=1 q=1

In order to calculate the unknown coefficients bi,j s, the collocation points defined
in Eqs. (3.1) and (3.2) are substituted in Eq. (3.3).Hence, we obtain the following
2M 2N linear system with unknowns bi,j s:
G(xm , yn ) =

2M
2N

p=1 q=1

bp,q hp (xm )hq (yn ), m = 1, 2, ..., 2M,

n = 1, 2, ..., 2N.

(10)

M. Fallahpour et al./ IJM2 C, 05 - 04 (2015) 361-372.

365

The solution of system (3.4) can be calculated from the following theorem.
Theorem 2. The solution of the system (3.4) is given below:

b1,1 =

2M
2N

1
G(xm , yn ),
2M 2N
p=1 q=1

bi,1 =

b1,j

1
2N

1 2N p=

G(xm , yn )

q=1

1
2N

G(xm , yn )

2
2
2M
2M

=
G(xm , yn )
G(xm , yn )
2M 2
q=
p=1

1
2
1

G(xm , yn )
1 2 p= q=
p=
1

i = 2, 3, ..., 2M,

j = 2, 3, ..., 2N,

p=1 q=2 +1

bi,j =

p=1 +1 q=1

G(xm , yn )

q=2 +1

G(xm , yn )

p=1 +1 q=2

G(xm , yn )

i = 2, 3, ..., 2M , j = 2, 3, ..., 2N,

p=1 +1 q=2 +1

where
1 = 1 (1 1) + 1,
1 = 1 (1 1) + 21 ,
1 = 1 1 ,
.
1 = 2M
1 ,
1 = i 1 ,
1 = 2log2 (i1)

(11)

2 = 2 (2 1) + 1,
2 = 2 (2 1) + 22 ,
2 = 2 2 ,
.
2 = 2N
2 ,
2 = j 2 ,
2 = 2log2 (j1)

(12)

and similarly,

Proof. See [2].

M. Fallahpour et al./ IJM2 C, 05 - 04 (2015) 361-372.

366

Consider a function G(x, y, s, t) of four variables x, y, s and t. Suppose G(x, y, s, t)

is approximated using two-dimensional Haar wavelet as follows [3]:

G(x, y, s, t)

2M
2N

bp,q (x, y)hp (s)hq (t).

(13)

p=1 q=1

Substituting the collocation points

si =

i 0.5
, i = 1, 2, ..., 2M,
2M

tj =

j 0.5
, j = 1, 2, ..., 2N,
2N

and

we obtain the linear system

G(x, y, si , tj )

2M
2N

bp,q (x, y)hp (si )hq (tj ) , i = 1, 2, ..., 2M , j = 1, 2, ..., 2N.

p=1 q=1

(14)
Corollary 1. The solution of the system (3.8) for any value of x.y [0, 1] is given
as follows [3]:
2M
2N

1
G(x, y, sp , tq ),
b1,1 (x, y) =
2M 2N
p=1 q=1

bi,1 (x, y) =

1
2N

1 2N p=

1
2N

G(x, y, sp , tq )

q=1

G(x, y, sp , tq ) , i = 2, 3, ..., 2M,

p=1 +1 q=1

2
2
2M
2M

G(x, y, sp , tq )
G(x, y, sp , tq ) , j = 2, 3, ..., 2N,
b1,j (x, y) =
2M 2
q=
p=1

p=1 q=2 +1

1
2
1

bi,j (x, y) =
G(x, y, sp , tq )
1 2 p= q=
p=
1

p=1 +1 q=2

G(x, y, sp , tq ) +

G(x, y, sp , tq )

q=2 +1

G(x, y, sp , tq ) ,

p=1 +1 q=2 +1

i = 2, 3, ..., 2M , j = 2, 3, ..., 2N,

M. Fallahpour et al./ IJM2 C, 05 - 04 (2015) 361-372.

367

where 1 , 1 , 1 and 1 are defined as in Eq. (3.5) and 2 , 2 , 2 and 2 are defined
as in Eq. (3.6).
Corollary 2. Suppose a function G(x, y) of two variables x and y is approximated
using Haar wavelet approximation given in Eq. (3.3). Suppose further that G(x, y)
is known at collocation points (xm , ym ), m = 1, 2, ..., 2M, n = 1, 2, ..., 2N. Then
the approximate value of the function G(x, y) at any other point of the domain
can be calculated as follows [3]:
2M
2N

1
G(x, y) =
G(xm , ym )h1 (x)h1 (y)
2M 2N
p=1 q=1

2M

i=1

2N

j=1

1
1
2N
2N

G(xm , ym )
G(xm , ym ) hi (x)h1 (y)
1 2N p=
1

2M 2

q=1

2
2M

p=1 +1 q=1

G(xm , ym )

p=1 q=2

2
2M

G(xm , ym ) h1 (x)hj (y)

p=1 q=2 +1

1
2
1
2M
2N

1
+
G(xm , ym )
1 2 p= q=
p=
i=1 j=1

G(xm , ym ) +

p=1 +1 q=2

F (xm , ym )

q=2 +1

G(xm , ym ) hi (x)hj (y),

p=1 +1 q=2 +1

where 1 , 1 , 1 and 1 are defined as in Eq. (3.5) and 2 , 2 , 2 and 2 are defined
as in Eq. (3.6).
3.2
Two-dimensional linear stochastic Fredholm integral equation
Consider the two-dimensional linear stochastic Fredholm integral equation (1.1).
First we define:
K2 (x, y, s, t)g(s, t) = (t, s),
afterward from (1.3) we have
s) = 1 {K2 (x, y, t, s)g(t, s) + K2 (x, y, s, t)g(s, t)} .
(t,
2
Assume that the function K(x, y, s, t)g(s, t) is approximated using two-dimensional
Haar wavelet as follows:
K1 (x, y, s, t)g(s, t)

2M
2N

i=1 j=1

bi,j (x, y)hi (s)hj (t),

(15)

M. Fallahpour et al./ IJM2 C, 05 - 04 (2015) 361-372.

368

s)
(t,

2M
2N

ci,j (x, y)hi (s)hj (t).

(16)

i=1 j=1

With this approximation and using Eq. (1.2) we can write Eq. (1.1) as follows:

2M
2N
1 1

g(x, y) = f (x, y) +
0

bi,j (x, y)hi (s)hj (t)dsdt

(17)

i=1 j=1

+2

t
2N
2M

ci,j (x, y)hi (s)hj (t)dB(s) dB(t).

0 i=1 j=1

Eq. (3.11) can be written in a more compact form using the notations introduced
in equations (2.1) and (2.2) and is given as follows:

g(x, y) = f (x, y) + b1,1 (x, y) + 2 c1,1 (x, y) Ito1(1) +

2N

c1,j (x, y) Ito2(j)

j=2

2M

ci,1 (x, y) Ito1(i) +

i=2

2M
2N

ci,j (x, y) Ito3(i, j) ,

i=2 j=2

where in recent equation from Eq. (1.4) we have:

Ito1(1) =

h1 (t)

h1 (t)
0

B 2 (1) 1
,
2
2

1 [ i

B(t)dB(t) =

Ito1(i) =

h1 (s)dB(s) dB(t) =

hi (s)dB(s) dB(t) =

dB(s)

]
dB(s) dB(t)

= (2B(i ) B(i ) B(i )) B(1),

Ito1(j) =

hj (t)
0

h1 (s)dB(s) dB(t) =
0

B(t)dB(t) =
j

and
Ito3(i, j) =

hj (t)
0

hj (t)B(t)dB(t)
0

B(t)dB(t)

]
hi (s)dB(s) dB(t)

2mB 2 (j ) B 2 (j ) B 2 (j ) + 1
,
2m

M. Fallahpour et al./ IJM2 C, 05 - 04 (2015) 361-372.

hj (t)
0

dB(s)

369

]
dB(s) dB(t)

= [2B(i ) B(i ) B(i )] [2B(j ) B(j ) B(j )] .

Substituting the collocation points given in (3.1) and (3.2), we obtain the following system of equations:

g(xm , yn ) = f (xm , yn ) + b1,1 (xm , yn ) + 2 c1,1 (xm , yn ) Ito1(1) +

2N

c1,j (xm , yn )

j=2

Ito2(j) +

2M

ci,1 (xm , yn ) Ito1(i) +

i=2

2M
2N

ci,j (xm , yn ) Ito3(i, j) .

i=2 j=2

Now bi,j , i = 1, 2, ..., 2M , j = 1, 2, ..., 2N and similarly ci,j , i =

1, 2, ..., 2M , j = 1, 2, ..., 2N can be replaced with their expressions given in
Corollary 1 and the following system of equations is obtained:
2M
2N

1
K1 (xm , yn , sp , tq )g(sp , tq )
2M 2N

g(xm , yn ) = f (xm , yn ) +

(18)

p=1 q=1

2N

j=2

2M

i=2

+2

M 2

2
2M

q , sp ) Ito1(1)
(t

p=1 q=1

q , sp )
(t

p=1 q=2

2
2M

q , sp )Ito2(j)
(t

p=1 q=2 +1

1
1
2N
2N

1
q , sp )
q , sp )Ito1(i)

(t
(t
1 N p=
1

2M
2N

i=2 j=2

1
M 2N

2M
2N

q=1

p=1 +1 q=1

1
2
1

(tq , sp )
1 2 p= q=
p=

p=1 +1 q=2

q , sp ) +
(t

p=1 +1 q=2 +1

q , sp )
(t

q=2 +1

q , sp ) Ito3(i, j).
(t

M. Fallahpour et al./ IJM2 C, 05 - 04 (2015) 361-372.

370
Table 1.

The solution mean with %95 confidence interval for above example

2M

(x, y)

g(x, y)
L

%95 Confidence
Interval

(0.25, 0.25)
(0.25, 0.75)
(0.75, 0.25)

0.394433
0.881874
0.882179

0.392491
0.879448
0.879495

0.396374
0.884299
0.884864

(0.125, 0.625)
(0.375, 0.375)
(0.875, 0.375)

0.626416
0.626752
1.06476

0.621964
0.619833
1.03276

0.630868
0.633671
1.09675

(0.0625, 0.8125)
(0.4375, 0.4375)
(0.9375, 0.1875)

0.625749
0.616593
0.854036

0.620075
0.611904
0.834678

0.631423
0.621282
0.873394

0.532988
0.459968
0.695789

0.522597
0.449544
0.686792

0.543379
0.470391
0.704786

0.0968562
0.87951
0.746029

0.0897225
0.868732
0.735251

0.10399
0.890288
0.756807

16

(0.03125, 0.84375)
(0.40625, 0.40625)
(0.9375, 0.1875)

16

32

(0.015625, 0.515625)
(0.765625, 0.765625)
(0.984375, 0.234375)

Eq. (3.12) represents 2M 2N system which can be solved using either prevalent
methods for solving linear systems. The solution of this system gives values of
g(x, y) at the collocation points. The values of g(x, y) at points other than
collocation points can be calculated using Corollary 2.

4.

Numerical Example

In this section, the numerical example is given to demonstrate the applicability and
accuracy of our method. Consider the following linear two dimensional stochastic
Fredholm integral equation of second kind:

u(x, y) = f (x, y)+

0

1 1

1 1

(x+y+ts)g(s, t)dsdt+
0

(x+y+t+s)g(s, t)dB(s)dB(t)
0

where
f (x, y) = x + y

1
xy(x3 + 4x2 y + 4xy 2 + y 3 ).
12

The solution mean with confidence interval at the collocation points 1000 iterative
of system (3.12) is shown in Table 1. In Figs. 1, three-dimensional graph of the
approximate solution for level L = 4 is shown.

M. Fallahpour et al./ IJM2 C, 05 - 04 (2015) 361-372.

371

Figure 1. Approximate solution plot (L = 4).

5.

Conclusion

The numerical solution of two-dimensional stochastic integral equations because

of the randomness is very difficult or sometimes impossible. In this paper, we have
successfully developed Haar wavelets numerical method for approximate a solution
of two-dimensional linear stochastic Fredholm integral equations. The example
confirm that the method is considerably fast and highly accurate as sometimes
lead to exact solution. Although, theoretically for getting higher accuracy we can
set the method with larger values of M and N and also larger of the degree of
approximation, p and q, but it leads to solving MN linear systems of size pq pq,
that have its difficulties. The method can be improved to be more accurate by
using other numerical methods. Mathematica has been used for computations.

References
[1] Ahmadi. S, Application of partial differential equations in snow mechanics, Mathematical Modelling
and Computations, (2011) 189-194.
[2] Aziz. I and Siraj-ul-Islam, New algorithms for numerical solution of nonlinear Fredholm and
Volterra integral equations using Haar wavelets, J. Comp. Appl. Math, 239 (2013) 333-345.
[3] Aziz. I, Siraj-ul-Islam and F. Khan, A new method based on Haar wavelet for numerical solution of
two-dimensional nonlinear integral equations, J. Comp. Appl. Math, 272 (2014) 70-80.
[4] Aleksandrov. V. M and Manzhirov. A. V, Two-dimentional integral equations in applied mechanics
of deformable solids, J. Appl. Mech. Tech. Phys, 5 (1987) 146-152.
[5] Assari. P, Adibi. H and Dehghal. M, A meshless method for solving nonlinear two-dimensional
integral equations of the second kind on non-rectangular domains using radial basis functions with
error analysis, J. Comp. Appl. Math, 239 (2013) 72-92.
[6] Atkinson. K. E, The numerical solution of integral equations of the second kind, Cambridge
University Press, (1997).
[7] Babolian. E, Maleknejad. K, Roodaki. M and Almasieh. H, Two dimensional triangular functions
and their applications to nonlinear 2d Volterra-Fredholm equations, Comp. Math. App, 60 (2010)
1711-1722.
[8] Bazm. S and Babolian. E, Numerical solution of nonlinear two-dimensional Fredholm integral
equations of the second kind using Gauss product quadrature rules, Commun. Nonlinear Sci.
Numer. Simult, 17 (2012) 1215-1223.
[9] Brunner. H, Collocation methods for Volterra integral and related functional equations, Cambridge
University Press, (2004).
[10] Dobner. H. J, Bounds for the solution of hyperbolic problems, Computing, 38 (1987) 209-218.
[11] Elkhayyari. O and Lamnii. A, Numerical solutions of second order boundary value problem by using
hyperbolic uniform b-splines of order 4, Mathematical Modelling and Computations, (2014) 25-36.
[12] Guoqiang. H, Itayami. H, Sugihara. K and Jiong. W, Extrapolation method of iterated collocation
solution for two-dimentional nonlinear Volterra integral equations, Appl. Math. Comput, 112
(2000) 49-61.
[13] Guoqiang. H and Jiong. W, Extrapolation of nystrom solution for two dimentional nonlinear
Fredholm integral equations, J. Comp. App. Math, 134 (2001) 259-268.
[14] Han. G and Wang. R, Richardson extrapolation of iterated discrete Galerkin solution for twodimensional Fredholm integral equations, J. Comp. App. Math, 139 (2002) 49-63.

372

M. Fallahpour et al./ IJM2 C, 05 - 04 (2015) 361-372.

[15] Hosseini Shekarabi. F, Maleknejad. K and Ezzati. R, Application of two-dimensional Bernstein

polynomials for solving mixed Volterra-Fredholm integral equations, African Mathematical Union
and Springer-Verlag Berlin Heidelberg, DOI 10. 1007/s 13370-014-0283-6, (2014).
[16] Jerri. A. J, Introduction to integral equations with applications, John Wiley and Sons, INC, (1999).
[17] Keinert. F, Wavelets and Multiwavelets, A Crc Press Company Boca Raton London New York
Washington, D. C, (2004).
[18] Kuo, Hui-Hsiung, Introduction to stochastic integration, Springer Science+Business Media, Inc,
(2006).
[19] Maleknejad. K and JafariBehbahani. Z, Application of two-dimensional triangular functions for
solving nonlinear class of mixed Volterra-Fredholm integral equations, Math. Comp. Mode, 55
(2012) 1833-1844.
[20] Manzhirov. A. V, Contact problems of the interaction between viscoelastic foundations subject to
ageing and systems of stamps not applied simultaneously, Prikl. Matem. Mekhan, 4 (1987) 523-535.
[21] Nemati. S and Ordokhani. Y, Solving nonlinear two-dimensional Volterra integral equations of the
first-kind using the bivariate shifted legendre functions, Mathematical Modelling and Computations,
(2015) 219- 230.
[22] Nemati. S, Lima. P and Ordokhani. Y, Numerical solution of a class of two-dimensional nonlinear
Volterra integral equations using legender polynomials, J. Comp. Appl. Math, 242 (2013) 53-69.
[23] Rahman. M, A rigid elliptical disc-inclusion in an elastic solid, subject to a polynomial normal
shift, J. Elasticity, 66 (2002) 207-235.
[24] Reihani. M. H and Abadi. Z, Rationalized Haar functions method for solving Fredholm and Volterra
integral equations, J. Comp. Appl. Math, 200 (2007) 12-20.
[25] Sankar. T. S and Fabrikant. V. I, Investigations of a two-dimentional integral equation in the theory
of elasticity and electrostatics, J. Mec. Theor. Appl, 2 (1983) 285-299.
[26] Siraj-ul-Islam, Aziz. I and Al-Fhaid. A, An improved method based on Haar wavelets for numerical
solution of nonlinear and integro-differential equations of first and higher orders, J. Comp. Appl.
Math, 260 (2014) 449-469.
[27] Siraj-ul-Islam, Aziz. I and Fayyaz. M, A new approach for numerical solution of integro-differential
equations via Haar wavelets, Int, J. Comp. Math, 90 (2013) 1971-1989.
[28] Soloviev. O. V, Low-frequency radio wave propagation in the earth-ionosphere waveguide disturbed
by a large-scale three-dimensional irregularity, Radiophysics and Quantum Electronics, 41 (1998)
392-402.
[29] Tari. A, Rahimi. M, Shahmorad. S and Talati. F, Solving a class of two-dimensional linear and
nonlinear Volterra integral equations by the differential transform method, J. Comp. Appl. Math,
228 (2009) 70-76.
[30] Xie. W and Lin. F. R, A fast numerical solution method for two dimensional Fredholm integral
equations of the second kind, App. Num. Math, 59 (2009) 1709-1719.