Applied Mathematics and Computation 256 (2015) 715723

Contents lists available at ScienceDirect

Applied Mathematics and Computation

journal homepage: www.elsevier.com/locate/amc

Legendre wavelets operational method for the numerical

solutions of nonlinear Volterra integro-differential equations
system
P.K. Sahu, S. Saha Ray
Department of Mathematics, National Institute of Technology Rourkela, Odisha 769008, India

a r t i c l e

i n f o

Keywords:
Legendre wavelets
Integro-differential equations
System of nonlinear Volterra integral
equations
Legendre wavelet method

a b s t r a c t
In this paper, Legendre wavelet method is developed to approximate the solutions of
system of nonlinear Volterra integro-differential equations. The properties of Legendre
wavelets are rst presented. The properties of Legendre wavelets are used to reduce the
system of integral equations to a system of algebraic equations which can be solved
numerically by Newtons method. Also, the results obtained by present method have been
compared with that of by B-spline wavelet method. Illustrative examples have been
discussed to demonstrate the validity and applicability of the present method.
2015 Elsevier Inc. All rights reserved.

1. Introduction
Wavelets theory is a relatively new and an emerging area in the eld of applied science and engineering. It has been
applied in a wide range of engineering disciplines; particularly, wavelets are very successfully used in signal analysis for
waveform representation and segmentation, timefrequency analysis and fast algorithms for easy implementation [1].
Wavelets permit the accurate representation of a variety of functions and operators. Moreover wavelets establish a connection with fast numerical algorithms [2]. Many researchers started using various wavelets for analyzing problems of high
computational complexity. It is proved that wavelets are powerful tools to explore new direction in solving differential equations and integral equations.
Integral equation has been one of the essential tools for various areas of applied mathematics. Integral equations occur
naturally in many elds of science and engineering [3]. Mathematical modeling of real-life problems usually results in functional equations, e.g. partial differential equations, integral and integro-differential equations, stochastic equations and others. Many mathematical formulations of physical phenomena contain integro-differential equations; these equations arise in
uid dynamics, biological models and chemical kinetics. Integro-differential equations arises in many physical processes,
such as glass-forming process [4], nano-hydrodynamics [5], drop wise condensation [6], wind ripple in the desert [7] and
biological model [8].
In the past several decades, many effective methods for obtaining approximation or numerical solutions of linear and
nonlinear integro-differential equations have been presented. There are various numerical and analytical methods to solve
such problems.
In the past, Legendre wavelet method (LWM) has been applied to solve the integral equations and integro-differential
equations of different forms. Recently, Mohamed and Torky [9] have solved system of linear Fredholm and Volterra integral
Corresponding author.
E-mail address: santanusaharay@yahoo.com (S. S. Ray).
http://dx.doi.org/10.1016/j.amc.2015.01.063
0096-3003/ 2015 Elsevier Inc. All rights reserved.

716

P.K. Sahu, S. S. Ray / Applied Mathematics and Computation 256 (2015) 715723

equation by applying Legendre wavelet method. In [10], the learned researchers Venkatesh et al. have applied Legendre
wavelet method to solve class of nonlinear integro-differential equations. Also the Legendre wavelet method has been
applied to solve nonlinear VolterraFredholm integral equations by Youse and Razzaghi in [11], and system of Fredholm
integral equations by Biazer and Ebrahimi in [12]. Biazer et al. in [13] have solved system of nonlinear Volterra integro-differential equations by using Homotopy perturbation method. In [14], system of linear Volterra integro-differential equations
has been solved by using Variational iteration method by Nadja and Tamamgar. Maleknejad et al. [15] have solved system
of high order linear Volterra Integro-differential equations by applying Bernstein operational matrix. Numerical methods for
solving Fredholm integral equations have been presented by learned researchers Saha Ray and Sahu in [16]. Linear semiorthogonal B-spline wavelets have been applied to solve integral equations and systems in [1618].
In this paper, we consider the system of nonlinear Volterra integro-differential equations of the following form
p

yi x Gi x; Yx

l Z
X

ki;j x; tF i;j t; Ytdt;

i 1; 2; . . . ; l:

1:1

j1
s

with initial conditions yi 0 bi;s ;

s 0; 1; . . . ; p  1;

i 1; 2; . . . ; l, where



1
p
1
p
Gi x; Yx Gi x; y1 x; y1 x; . . . ; y1 x; . . . ; yl x; yl x; . . . ; yl x ;


1
p
1
p
F i;j t; Yt F i;j t; y1 t; y1 t; . . . ; y1 t; . . . ; yl t; yl t; . . . ; yl t ;
p

and ki;j x; t are the kernel functions for i; j 1; 2; . . . ; l and yi x is the pth order derivative of yi x for i 1; 2; . . . ; l.
In this present paper, we apply Legendre wavelet method (LWM) to solve the system of nonlinear Volterra integro-differential equations. The Legendre wavelet method converts the system of integro-differential equations to a system of algebraic equations and that algebraic equations system again can be solved by any of the usual numerical methods. The results
obtained by present method have been compared with the results obtained by B-spline wavelet method (BWM). In aid to
compare, the same solution method has been implemented to both the methods.
2. Properties of Legendre wavelets
Wavelets constitute a family of functions constructed from dilation and translation of a single function called mother
wavelet. When the dilation parameter a and the translation parameter b vary continuously, we have the following family
of continuous wavelets as
1



xb
;
a

Wa;b x j aj2 W

a; b 2 R;

a0

2:1

k
If we restrict the parameters a and b to discrete values as a ak
0 ; b nb0 a0 ; a0 > 1; b0 > 0 and n, and k are positive integers, from Eq. (2.1) we have the following family of discrete wavelets:
k

wk;n x j a0 j2 wak0 x  nb0 ;

where wk;n x form a wavelet basis for L2 R. In particular, when a0 2 and b0 1, then wk;n x form an orthonormal basis.
^ ; m; x have four arguments. n
^ 2n  1; n 1; 2; . . . ; 2k1 ; k 2 Z ; m is the order of
Legendre wavelets wn;m x wk; n
Legendre polynomials and x is normalized time. They are dened on 0; 1 as

^ ; m; x
wn;m x wk; n

( q k
^ ;
m 1222 Pm 2k x  n
0;

^ 1
n
2k

6 x < n^21
k

2:2

otherwise

where m 0; 1; . . . ; M  1 and n 1; 2; . . . ; 2k1 . The coefcient

q
m 12 is for orthonormality, the dilation parameter is

^ 2k .
a 2k and translation parameter is b n
Here P m x is Legendre polynomials of order m, which are orthogonal with respect to weight function wx 1 on the
interval 1; 1. This can be determined from the following recurrence formulae

P0 x 1;
P1 x x;
Pm1 x





2m 1
m
xPm x 
Pm1 x;
m1
m1

m 1; 2; 3; . . .

3. Function approximation by Legendre wavelets

A function f x dened over 0; 1 can be expressed by the Legendre wavelets as

P.K. Sahu, S. S. Ray / Applied Mathematics and Computation 256 (2015) 715723

f x

1 X
1
X
cn;m wn;m x

717

3:1

n1 m0

where cn;m hf x; wn;m xi, in which h:; :i denotes the inner product. If the innite series in Eq. (3.1) is truncated, then Eq.
(3.1) can be written as

f x

2k1 M1
X
X

cn;m wn;m x C T Wx;

3:2

n1 m0



where C and Wx are 2k1 M  1 matrices given by

h
iT
C c1;0 ; c1;1 ; . . . ; c1;M1 ; c2;0 ; . . . ; c2;M1 ; . . . ; c2k1 ;0 ; . . . ; c2k1 ;M1 ;
h

3:3
iT

Wx w1;0 x; w1;1 x; . . . ; w1;M1 x; . . . ; w2k1 ;0 x; . . . ; w2k1 ;M1 x :

3:4

3.1. Operational matrix of derivative [19]

The operational matrix of derivative for Legendre wavelets D can be dened as follow

W0 x DWx
W00 x D2 Wx
:
:
:

Wp x Dp Wx
where

H
6
60
6
D6
60
6
4:

7
07
7
: 07
7
7
: :5
: H

3:5

in which H is M 1  M 1 matrix and its each element is dened as follow

(
Hr;s

2k1

p
2r  12s  1; r 2; . . . ; M 1; s 1; . . . ; r  1;

0;

and r s odd

otherwise

3.2. Operational matrix of integration [20]

The operational matrix of integration for Legendre wavelets Q can be dened as follow

Wxdx Q Wx

Z0 x Z
0

Wxdxdx Q 2 Wx

:
:
:
Z

...
0

Wxdxdx
. . . dx Q p Wx
|{z}
ptimes


where Q is a 2k1 M  2k1 M matrix and dened as follow

718

P.K. Sahu, S. S. Ray / Applied Mathematics and Computation 256 (2015) 715723

: W

6
60
16
Q k6
60
2 6
4:

7
: W7
7
: W7
7
7
: : 5
: X

3:6

where X and W are M  M matrices dened as

2 0 0 : 0

7
6
60 0 0 : 07
7
6
7
W 6
60 0 0 : 07
7
6
4: : : : :5
0 0 0 : 0
and

p1
3

6 1
6  p
0
3
6
6
6 0
 p31p5
6
6 :
:
6
X6
6 :
:
6
6 :
:
6
6
6 0
0
4
0

: :

p1p
3 5

: :

p1p
5 7

: :

: :

: :

7
7
7
7
7
0
0
7
7
:
:
7
7
7
:
:
7
7
:
:
7
7
1p 7
p

0
2M3 2M1 5
1p
 p
0
2M1 2M3
0

:
:

4. Legendre wavelet method for system of nonlinear Volterra integro-differential equations

We consider the system of nonlinear Volterra integro-differential equations given in Eq. (1.1). First, we approximate the
unknown functions by Legendre wavelet method using Eq. (3.2) as

yi x ATi Wx;

i 1; 2; . . . ; l;

4:1

where ATi is similar to C T dened in Eq. (3.3) and Wx is dened in Eq. (3.4).
Rx Rx
Rx
1
dp
Let us dene L  dx
 0 0 . . . 0 : dxdx
. . . dx. we can also approximate the derivative and integration of
p ;L
|{z}
ptimes

yi x

i 1; 2; . . . ; l by operational matrices of Legendre wavelets as

yi x ATi W1 x ATi DWx

1

yi x ATi W2 x ATi D2 Wx
2

:
:
:
yi x ATi Wp x ATi Dp Wx;
p

and

i 1; 2; . . . ; l;

Z x
yi xdx ATi
Wxdx ATi Q Wx
0
0
Z xZ x
Z xZ x
yi xdxdx ATi
Wxdxdx ATi Q 2 Wx
x

:
:
:
Z

...
0

yi xdxdx
. . . dx ATi
|{z}
ptimes

...
0

Wxdxdx
. . . dx ATi Q p Wx
|{z}
ptimes

i 1; 2; . . . ; l;
where D and Q are dened in Eqs. (3.5) and (3.6), respectively.

P.K. Sahu, S. S. Ray / Applied Mathematics and Computation 256 (2015) 715723

719

Now substituting Eq. (4.1) in Eq. (1.1) and then operating L1 both sides, we have
1

L L

ATi

"
l Z
i
X
1
1
Wx L Gi x; Xx L

ki;j x; tF i;j t; Xtdt ; i 1; 2; . . . ; l;

4:2

j1

where

Xx x; AT1 Wx; AT1 D1 Wx . . . ; AT1 Dp Wx; . . . ; ATl Wx; ATl D1 Wx; . . . ; ATl Dp Wx:
We set

"
U i x L1 Gi x; Xx;

V i x L1

l Z
X
j1

ki;j x; tF i;j t; Xtdt

The Eq. (4.2) can be reduced as

ATi Wx

p1
X
bi;s xs U i x V i x;

i 1; 2; . . . ; l;

4:3

s0

where

bi;s yi 0 ATi Ds W0;

s

i 1; 2; . . . ; l;

Utilizing Eq. (4.3) by collocation points xr 2r1

;
2k M
with same number of unknowns for Ai ;

s 0; 1; . . . ; p  1:
r 1; 2; . . . ; 2k1 M, we get a algebraic system of 2k1 Ml equations along

i 1; 2; . . . ; l. Solving this system numerically by Newtons method, we can get the

i 1; 2; . . . ; l and hence obtain the solutions yi x ATi Wx;

value of unknowns for Ai ;

i 1; 2; . . . ; l.

5. Convergence analysis
Theorem 1. The series solution yx
yx.

P1 P1
n1

m0 c n;m wn;m x

dened in Eq. (3.1) using Legendre wavelet method converges to

Proof. Let L2 R be the Hilbert space and wn;m dened in (2.2) forms an orthonormal basis.
P
Let yx M1
i0 C n;i wn;i x where C n;i hyx; wn;i xi for a xed n.
Let us denote wn;i x wx and let aj hyx; wxi.


Now we dene the sequence of partial sums Sn of aj wxj ; Let Sn and Sm be the partial sums with n P m. We have to
prove Sn is a Cauchy sequence in Hilbert space.
P
Let Sn ni1 aj wxj .
Now

hyx; Sn i

yx;

n
X

aj wxj

n
X

i1

j aj j2 :

j1

We claim that

kSn  Sm k2

n
X

j aj j2 ;

n > m:

jm1

Now

2 *

X
n
n
n
n
X
X
X

aj wxj

aj wxj ;
aj wxj
j aj j2 ;

jm1
jm1
jm1
jm1

for

Therefore,

X
n
n
X

aj wxj

j a j j2 ;

jm1
j1
From Bessels inequality, we have

for
Pn

j1

n > m:
j aj j2 is convergent and hence

X
n

aj wxj

! 0 as m; n ! 1

jm1

n > m:

720

P.K. Sahu, S. S. Ray / Applied Mathematics and Computation 256 (2015) 715723

So;

X
n

a
wx

j
j
! 0

jm1

and fSn g is a Cauchy sequence and it converges to s (say).

We assert that yx s.
Now

D
E
hs  yx; wxj i hs; wxj i  hyx; wxj i lim Sn ; wxj  aj aj  aj
n!1

This implies

hs  yx; wxj i 0
Hence yx s and

Pn

j1

aj wxj converges to yx as n ! 1 and proved.

6. Illustrative examples
Example 1. Consider the following system of Volterra integro-differential equations [13]

1
1
1
y001 x 1  x3  y02
x
3
2 2
2
y002 x 1 x2  xy1 x

1
4

Z
0

y21 t y22 t dt

y21 t  y22 t dt

with initial conditions are y1 0 1; y01 0 2; y2 0 1; y02 0 0. The exact solutions are y1 x x e x and
y2 x x  e x . The approximate solutions obtained by Legendre wavelet method for M 8 and k 2 with their exact solutions have been shown in Table 1 and the comparison of absolute errors obtained by Legendre wavelet method and B-spline
wavelet method (m 4) have been presented in Table 2. L2 error for y1 x obtained by LWM and BWM are 1.09139E-10 and
2.03738E-4, respectively. Similarly, L2 error for y2 x obtained by LWM and BWM are 1.10249E-10 and 2.07909E-4, respectively. The over all computational time is recorded as 1118.65 s for LWM and 30971.20 s for BWM for this problem. Clearly, it
can be observed that LWM is more accurate than BWM with regards to absolute errors as well as L2 errors. Furthermore, the
computational time in LWM is much less than BWM.
Example 2. Consider the following system of Volterra integro-differential equations [13]

Z x
1
y01 x 1  y02
x

x  ty2 t y2 ty1 tdt

2 2
0
Z x


y02 x 2x
x  ty1 t  y22 t y21 t dt;
0

with the initial conditions y1 0 0; y2 0 1. The exact solutions are y1 x sinh x and y2 x cosh x. The approximate
solutions obtained by Legendre wavelet method for M 4 and k 2 with their exact solutions have been cited in Table 3 and
the comparison of absolute errors obtained by Legendre wavelet method and B-spline wavelet method (m 2) have been
presented in Table 4. L2 error for y1 x obtained by LWM and BWM are 7.25142E-5 and 1.32592E-3, respectively. Similarly,
L2 error for y2 x obtained by LWM and BWM are 1.10378E-4 and 2.73105E-3, respectively. The over all computational time
Table 1
Numerical results obtained by Legendre wavelet method with their exact results for Example 1.
x

0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9

Numerical solutions

Exact solutions

y1 x

y2 x

y1 x

y2 x

1
1.20517
1.4214
1.64986
1.89182
2.14872
2.42212
2.71375
3.02554
3.3596

1
1.00517
1.0214
1.04986
1.09182
1.14872
1.22212
1.31375
1.42554
1.5596

1
1.20517
1.4214
1.64986
1.89182
2.14872
2.42212
2.71375
3.02554
3.3596

1
1.00517
1.0214
1.04986
1.09182
1.14872
1.22212
1.31375
1.42554
1.5596

721

P.K. Sahu, S. S. Ray / Applied Mathematics and Computation 256 (2015) 715723
Table 2
Absolute errors obtained by Legendre wavelet method and B-spline wavelet method for Example 1.
x

0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9

Legendre wavelet method

B-spline wavelet method

y1 x

y2 x

y1 x

y2 x

5.69869E-11
5.07594E-13
1.7697E-13
1.0747E-13
7.38298E-13
9.2065E-11
2.81153E-12
4.16023E-12
6.79279E-12
1.07501E-11

5.71666E-11
4.49196E-13
3.01981E-13
3.05089E-13
4.65183E-13
9.42553E-11
8.09797E-13
3.58158E-13
2.73115E-13
1.12377E-12

1.22179E-4
4.31693E-5
1.65916E-5
4.85740E-5
4.19752E-5
1.68426E-5
3.66785E-5
4.69159E-5
1.58263E-6
1.28407E-4

1.22177E-4
4.29516E-5
1.76404E-5
5.16821E-5
4.96080E-5
4.02196E-8
6.43960E-5
8.14876E-5
3.49124E-5
9.48954E-5

is recorded as 424.151 s for LWM and 1796.63 s for BWM for this problem. Clearly, it can be observed that LWM is more
accurate than BWM with regards to absolute errors as well as L2 errors. Furthermore, the computational time in LWM is
much less than BWM.
Example 3. Consider the following system of Volterra integro-differential equations [3]

1
1
1
2
y001 x cosh x  sinh x  x4  x2
2
6
2

y002 x 1 4x cosh x 8 sinh x  4x

x  ty21 t x  ty22 t dt

x
0

x  ty21 t  x  ty22 t dt;

with the initial conditions y1 0 1; y01 0 1 and y2 0 1 y02 0 1. The exact solutions are y1 x x cosh x and
y2 x x  cosh x. The approximate solutions obtained by Legendre wavelet method for M 8 and k 2 with their exact
solutions have been shown in Table 5 and the comparison of absolute errors obtained by Legendre wavelet method and
B-spline wavelet method (m 4) have been presented in Table 6. L2 error for y1 x obtained by LWM and BWM are
Table 3
Numerical results obtained by Legendre wavelet method with their exact results for Example 2.
x

0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9

Numerical solutions

Exact solutions

y1 x

y2 x

y1 x

y2 x

1.69785E-5
0.100168
0.201334
0.30458
0.410754
0.521029
0.636648
0.758569
0.88809
1.02651

0.999932
1.00501
1.02007
1.04534
1.08108
1.12754
1.18548
1.25517
1.33743
1.43310

0
0.100167
0.201336
0.30452
0.410752
0.521095
0.636654
0.758584
0.888106
1.02652

1
1.005
1.02007
1.04534
1.08107
1.12763
1.18547
1.25517
1.33743
1.43309

Table 4
Absolute errors obtained by Legendre wavelet method and B-spline wavelet method for Example 2.
x

0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9

Legendre wavelet method

B-spline wavelet method

y1 x

y2 x

y1 x

y2 x

1.69785E-5
9.28988E-7
1.57109E-6
2.17615E-6
1.24928E-6
6.63413E-5
5.12473E-6
1.48949E-5
1.64184E-5
6.44674E-6

6.82587E-5
1.03030E-5
1.49865E-6
1.31948E-6
1.10079E-5
8.30807E-5
1.38480E-5
8.59612E-7
9.67972E-7
1.40323E-5

2.08424E-9
2.65640E-4
2.84298E-4
2.11215E-4
1.96044E-4
3.93187E-4
2.39372E-5
1.85522E-4
1.23009E-4
1.14844E-3

1.98123E-3
6.34799E-4
2.57898E-4
7.05019E-4
6.67974E-4
5.88992E-5
6.76305E-4
7.83386E-4
2.82383E-4
9.83028E-4

722

P.K. Sahu, S. S. Ray / Applied Mathematics and Computation 256 (2015) 715723

Table 5
Numerical results obtained by Legendre wavelet method with their exact results for Example 3.
x

0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9

Numerical solutions

Exact solutions

y1 x

y2 x

y1 x

y2 x

1
1.105
1.22007
1.34534
1.48107
1.62763
1.78547
1.95517
2.13743
2.33309

1
0.905004
0.820067
0.745339
0.681072
0.627626
0.585465
0.555169
0.537435
0.533086

1
1.105
1.22007
1.34534
1.48107
1.62763
1.78547
1.95517
2.13743
2.33309

1
0.905004
0.820067
0.745339
0.681072
0.627626
0.585465
0.555169
0.537435
0.533086

Table 6
Absolute errors obtained by Legendre wavelet method and B-spline wavelet method for Example 3.
x

0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9

Legendre wavelet method

B-spline wavelet method

y1 x

y2 x

y1 x

y2 x

4.69155E-11
3.6593E-13
2.40918E-13
2.367E-13
3.85914E-13
5.82778E-11
5.1581E-13
2.05169E-13
1.47882E-13
7.10543E-13

4.69159E-11
3.65152E-13
2.43361E-13
2.44027E-13
3.68039E-13
5.82324E-11
4.55191E-13
3.06866E-13
3.16414E-13
4.46421E-13

1.22177E-4
3.92678E-5
1.49084E-5
4.06007E-5
3.64909E-5
3.42951E-7
4.04541E-5
4.86185E-5
1.83757E-5
5.80461E-5

1.22177E-4
3.92671E-5
1.49195E-5
4.06552E-5
3.66545E-5
2.43668E-8
4.11247E-5
4.96506E-5
1.97573E-5
5.64004E-5

7.48238E-11 and 1.65465E-4, respectively. Similarly, L2 error for y2 x obtained by LWM and BWM are 7.47871E-11 and
1.65577E-4, respectively. The over all computational time is recorded as 1512.60 s for LWM and 40879.60 s for BWM for
this problem. Clearly, it can be observed that LWM is more accurate than BWM with regards to absolute errors as well as
L2 errors. Furthermore, the computational time in LWM is much less than BWM.
7. Conclusion
In this present work, Legendre wavelet method has been applied to solve the system of nonlinear Volterra integro-differential equations. Using this procedure, the system of integro-differential equations have been reduced to solve a system of
algebraic equations. The obtained numerical results highly agree with the exact results. Additionally, the results obtained by
present method have been compared with the that of by B-spline wavelet method. Since, the same solution method has been
implemented to both the methods, from the tables it manifests that the present method solutions are better than other
method solutions. More over, the computational time of present method is very less than that of by B-spline wavelet method.
The illustrative examples have been included to demonstrate the validity and applicability of the present technique. These
examples also exhibit the accuracy and efciency of the present method.
References
[1] C.K. Chui, Wavelets, a mathematical tool for signal analysis, in: SIAM Monographs on Mathematical Modeling and Computation, SIAM, Pennsylvania,
1997.
[2] G. Beylkin, R. Coifman, V. Rokhlin, Fast wavelet transforms and numerical algorithms, I. Commun. Pure Appl. Math. 44 (1991) 141183.
[3] A.M. Wazwaz, Linear and Nonlinear Integral Equations: Methods and Applications, Springer-Verlag, Beijing, 2011.
[4] H. Wang, H.M. Fu, H.F. Zhang, Z.Q. Hu, A practical thermodynamic method to calculate the best glass-forming composition for bulk metallic glasses, Int.
J. Nonlinear Sci. Numer. Simul. 8 (2) (2007) 171178.
[5] L. Xu, J.H. He, Y. Liu, Electrospun nanoporous spheres with Chinese drug, Int. J. Nonlinear Sci. Numer. Simul. 8 (2) (2007) 199202.
[6] F.Z. Sun, M. Gao, S.H. Lei, Y.B. Zhao, K. Wang, Y.T. Shi, N.H. Wang, The fractal dimension of the fractal model of dropwise condensation and its
experimental study, Int. J. Nonlinear Sci. Numer. Simul. 8 (2) (2007) 211222.
[7] T.L. Bo, L. Xie, X.J. Zheng, Numerical approach to wind ripple in desert, Int. J. Nonlinear Sci. Numer. Simul. 8 (2) (2007) 223228.
[8] F. Shakeri, M. Dehghan, Solution of a model describing biological species living together using the variational iteration method, Math. Comput. Model.
48 (2008) 685699.
[9] M.A. Mohamed, M.Sh. Torky, Legendre wavelet for solving linear system of Fredholm and Volterra integral equations, Int. J. Res. Eng. Sci. 1 (7) (2013)
1422.

P.K. Sahu, S. S. Ray / Applied Mathematics and Computation 256 (2015) 715723

723

[10] S.G. Venkatesh, S.K. Ayyaswamy, S.R. Balachandar, K. Kannan, Wavelet solution for class of nonlinear integro-differential equations, Indian J. Sci.
Technol. 6 (2013) 46704677.
[11] S. Youse, M. Razzaghi, Legendre wavelets method for the nonlinear VolterraFredholm integral equations, Math. Comput. Simul. 70 (2005) 18.
[12] J. Biazar, H. Ebrahimi, Legendre wavelets for systems of Fredholm integral equations of the second kind, World Appl. Sci. J. 9 (9) (2010) 10081012.
[13] J. Biazar, H. Ghazvini, M. Eslami, Hes homotopy perturbation method for systems of integro-differential equations, Chaos, Solitons Fractals 39 (2009)
12531258.
[14] J.S. Nadja, M. Tamamgar, The variational iteration method: a highly promising method for solving the system of integro-differential equations,
Comput. Math. Appl. 56 (2008) 346351.
[15] K. Maleknejad, B. Basirat, E. Hashemizadeh, A Bernstein operational matrix approach for solving a system of high order linear Volterra Fredholm
integro-differential equations, Math. Comput. Model. 55 (2012) 13631372.
[16] S.S. Ray, P.K. Sahu, Numerical methods for solving Fredholm integral equations of second kind, Abstr. Appl. Anal. 2013 (2013) 17. Article ID. 426916.
[17] P.K. Sahu, S.S. Ray, Numerical solutions for the system of Fredholm integral equations of second kind by a new approach involving semiorthogonal Bspline wavelet collocation method, Appl. Math. Comput. 234 (2014) 368379.
[18] P.K. Sahu, S.S. Ray, A new approach based on semi-orthogonal B-spline wavelets for the numerical solutions of the system of nonlinear Fredholm
integral equations of second kind, Comput. Appl. Math. (2013), http://dxdoi.org/10.1007/s40314-013-0100-0.
[19] F. Mohammadi, M.M. Hosseini, A new Legendre wavelet operational matrix of derivative and its applications in solving the singular ordinary
differential equations, J. Franklin Inst. 348 (2011) 17871796.
[20] M. Razzaghi, S. Youse, The Legendre wavelets operational matrix of integration, Int. J. Syst. Sci. 32 (4) (2001) 495502.