Available at Applications and Applied

http://pvamu.edu/aam
Appl. Appl. Math.
Mathematics:
ISSN: 1932-9466 An International Journal
(AAM)
Vol. 6, Issue 1 (June 2011) pp. 268 – 283
(Previously, Vol. 6, Issue 11, pp. 2009 – 2024)

Homotopy Perturbation Method for Solving System of

Generalized Abel’s Integral Equations

Sunil Kumar, Om P. Singh* and Sandeep Dixit

Department of Applied Mathematics
Institute of Technology
Banaras Hindu University
Varanasi -221005, India
[email protected], [email protected],
[email protected]

Received: July 10, 2010; Accepted: February 3, 2011

*
Corresponding author

Abstract
In this paper, a user friendly algorithm based on the homotopy perturbation method (HPM) is
proposed to solve a system of generalized Abel’s integral equations. The stability of the solution
under the influence of noise in the input data is analyzed. It is observed that the approximate
solutions converge to the exact solutions. Illustrative numerical examples are given to
demonstrate the efficiency and simplicity of the proposed method in solving such types of
systems of Abel’s integral equations.

Keywords: Homotopy Perturbation Method, System of Generalized Abel’s Integral

Equations, Abel’s Kernel.

MSC (2010) No.: 45E10, 45F15

268
AAM: Intern. J., Vol. 6, Issue 1 (June 2011) [Previously, Vol. 6, Issue 11, pp. 2009 – 2024] 269

1. Introduction

The Volterra integral equations were studied by Traian Lalescu in his thesis, Sur les equations de
Volterra, written under the supervision of Emile Picard. Lalescu (1912) wrote the first book ever
on integral equations. These equations find applications in demography, the study of viscoelastic
materials, and in insurance mathematics through renewal equation. Volterra integral equations
arise in many problems pertaining to mathematical physics like heat conduction problems. Many
interesting problems of mechanics and physics lead to an integral equation in which the kernel
K ( x, t ) is of convolution type, that is, K ( x, t )  k ( x  t ), where k is a certain function of one
variable. Recently, systems of Volterra integral equations of convolution type have attracted the
attention of many authors and several different methods have been developed to solve these
numerically. In particular, we consider the following system of singular Volterra integral
equation of convolution type

x
c( x) y ( x)    k ( x  t ) a y (t ) dt  f ( x) , 0  x  1, (1)
0

where c( x)  (ci j ( x))i j , a  (ai j ) i j are square matrices of order n ,

y ( x)   y1 ( x), y 2 ( x), , y n ( x) T and f ( x)   f1 ( x), f 2 ( x),  , f n ( x)

T
(2)

are column vectors and   (i i ) i i and k ( x  t )  (ki i ( x  t ))i i are diagonal matrices of order n .

In equation (1), the functions k and f are given and y is the vector function of the solution of
the system (1) to be determined. Here, we assume that the system (1) has a unique solution. If the
domain of definition of the kernel is infinite, or if the kernel has a singularity within its domain
of definition, then the integral equation is said to be singular. In certain cases, the kernel is only
weakly singular as the singularity may be transformed away by a change of variable. In the
Volterra integral equations system of convolution type, if at least one of the integral equations is
singular, then the system is called system of singular Volterra integral equations of convolution
type. There are several numerical method for solving equation (1) for example, Galerkin,
collocation, Taylor series and Taylor polynomials Methods, [Burton (2005), Maleknejad and
Kajani (2004), Maleknejad and Aghazadeh (2005) and Yalsinbas (2002)]. Recently a number of
algorithms have been proposed based on Legendre wavelets [Malekkejad and Kajani (2003)]
Chebyshev polynomials [Malekkejad et al. (2007)], Bernstein polynomials [Pandey et al.
(2009)], Expansion method [Rabbani (2007)], Power Series method [Tahmasbi and Fard (2008)]
and almost Bernstein operational matrix method [Singh et al. (2010)].
1
For the special case when the kernel k ( x  t )  , where is a diagonal matrix of order n
( x  t )
with all the entries lying in (0, 1): the system (1) reduces to a system of generalized Abel’s
integral equations.

In this paper, we have developed a simple algorithm based on homotopy perturbation method for
the numerical solution of system of generalized Abel’s integral equations.
270 Sunil Kumar et al.

2. Basic Idea of Homotopy Perturbation Method

In this method, using the homotopy technique of topology, a homotopy is constructed with an
embedding parameter p  0 ,1  , which is considered as a “small parameter”. This method
became very popular amongst the scientists and engineers, even though it involves continuous
deformation of a simple problem into a more difficult problem under consideration. Most of the
perturbation methods depend on the existence of a small perturbation parameter but many
nonlinear problems have no small perturbation parameter at all. Many new methods have been
proposed in the late nineties to solve such nonlinear equation devoid of such small parameters
[Dehghan and Shakeri (2008, 2008), Ganji and Rajabi (2006), He (1999, 1999), Lio (1999,
1997)]. Late 1990s saw a surge in applications of homotopy theory in the scientific and
engineering computations [Abbasbandy (2006, 2007), Aminikhah and Salahi (2009), Biazar et
al. (2009), Biazar and Eslami (2010)].When the homotopy theory is coupled with perturbation
theory it provides a powerful mathematical tool [Ganji et al. (2007), He (1998, 2004), Shakeri
and Dehghan (2008)]. A review of recently developed methods of nonlinear analysis can be
found in He (2000). To illustrate the basic concept of HPM, consider the following nonlinear
functional equation

 u 
A ( u )  f ( r ), r   , with the boundary conditions: B  u ,   0, r   , (3)
 n 

where A is a general functional operator, B is a boundary operator, f (r ) is a known analytic

function, and   is the boundary of the domain  . The operator A is decomposed as A  L  N ,
where L is the linear and N is the nonlinear operator. Hence, Equation (3) can be written as

L (u )  N (u )  f ( r )  0 , r  .

We construct a homotopy  (r , p) :   0,1  R, satisfying

H ( , p )  (1  p ) L ( )  L ( u 0 )   p  A ( )  f ( r )   0 , p  [ 0 ,1], r   . (4)

Hence,

H ( , p)  L( )  L(u0 )  pL(u0 )  pN (v)  f (r )  0, (5)

where u 0 is an initial approximation for the solution of equation (3). As

H ( v ,0 )  L ( v )  L (u 0 ) and H(v,1)  A(v)  f (r),

AAM: Intern. J., Vol. 6, Issue 1 (June 2011) [Previously, Vol. 6, Issue 11, pp. 2009 – 2024] 271

it shows that H ( v , p ) continuously traces an implicitly defined curve from a starting point
H (u0 ,0) to a solution H (v,1) .The embedding parameter p increases monotonously from zero to
one as the trivial linear part L (u )  0 deforms continuously to the original problem
A(u )  f ( r ). The embedding parameter p  0 ,1  can be considered as an expanding parameter
[He (1999)] to obtain

v  v 0  pv 1  p 2 v 2  ... (6)

The solution is obtained by taking the limit as p tends to 1 in equation (6). Hence,

u  lim v  v0  v1  v2   . (7)
p 1

The series ( 7 ) converges for most cases and the rate of convergence depends on A ( u )  f ( r ) , He
(1999).

3. Method of Solution

We consider the following system of generalized Abel’s integral equation

x
a y (t )
cy ( x )  f ( x )    dt , 0  x  1, (8)
0
( x  t )

which is equivalent to the following set of n equations

x n a ij y j ( t )
ci y i ( x )  f i ( x )  i   
dt , 1  i  n. (9)
0 j 1 (x  t) jj

To solve the above equation, a convex homotopy is constructed as

 x n
a ij L ji (t ) 
(1  p )Li ( x )  y i 0 ( x )   p  Li ( x )  f i ( x )   i    jj
dt   0 . (10)
 0 j 1 ( x  t ) 

We seek the solution of (10) in the following form,


Li ( x )  p
j 0
j
Lij ( x ) , i  1, 2 ,3,... n , (11)
272 Sunil Kumar et al.

where L ij ( x ), i , j  1, 2 ,3,... , are the functions to be determined. After choosing the initial
approximations L i 0 ( x )  y i 0 ( x ) appropriately for 1  i  n , the following iterative scheme is
used to evaluate L ij ( x ).
Substituting equation (11) in (10) and equating the coefficients of p with the same power, we
get
x
 n a L (t ) 
p1 : L i1 ( x )   i    i j j 0  ij  dt ,
0  j 1 ( x  t ) 
x
 n a i j L j1 ( t ) 
p2 : L i 2 ( x )   i    
 ij  dt ,
0  j 1 ( x  t ) 
x
 n a L (t ) 
p3 : L i 3 ( x )   i    i j j 2  ij  dt ,
0  j 1 ( x  t ) 
.
.
.
x
 n a L (t ) 
p m
: L im ( x )   i    i j j ( m 1) ij  dt , (12)
0  j 1 (x  t) 

Hence, the solution vector y of equation (8) is given by y  ( y1 , y 2 , y 3 ,... y n ), where

m
y i ( x )  lim Li ( x )  lim
p 1 m

j0
Lij ( x ). (13)

4. Stability Analysis

We consider the stability of the solution components yi as given by (13) under the influence of
noise in the input data f (x ). That is, we wish to investigate the effect on the solution y ( x ) when
the input f (x ) is corrupted with noise f (x ) where f (x ) is unknown apart from some
restriction on its magnitude relative to f (x ).

Assuming ~
y ( x ) to be the solution of equation (8) under the influence of noise

f ( x )  f1 ( x ), f 2 ( x ), f 3 ( x ),..., f n ( x ) ,

then

x a~ y (t )
y ( x)  f ( x)   f ( x)   
~ dt , 0  x  1. (14)
0 ( x  t )
AAM: Intern. J., Vol. 6, Issue 1 (June 2011) [Previously, Vol. 6, Issue 11, pp. 2009 – 2024] 273

~
Choosing the initial approximation L0 ( x)  y0 ( x)   0 ( x)  L0   0 , where

~

~ ~ ~ ~

L0  L10 , L20 , L30 ,..., Ln 0 , y0 ( x )   y10 , y 20 , y30 ,..., y n 0  and  0 ( x)   10 ,  20 ,  30 ,...,  n 0   f ( x),

we have, by equation (12)

~ x a  y (t )   (t ) 
L1 ( x)    0 0
dt  L1 ( x)   1 ( x),
0 ( x  t )

~
L2 ( x)  L2 ( x)   2 ( x),
.
.
.

Lm ( x)  Lm ( x)   m ( x),

where the various components of Lm (x ) are given by equation (12) and

x a m 1 (t )
 m ( x)    dt , m  1, 2,3,.
0 ( x  t )
m
~
Thus, the perturbed solution ~y ( x ) is given by ~y ( x )  lim  L j ( x ).
m
j 0

The effect of the noise term  0 ( x ) in the input data f (x ) deviates the solution by

m n
 y ( x )  y ( x )  y ( x )  lim   Li ( x )  Li ( x )   lim   i ( x ). (15)
m  n 
i 0 i 0

From equation (15), we conclude that y (x ) and f (x ) are connected via the following
generalized Abel integral equation

x y (t )
cy ( x)  f ( x)    dt. (16)
0 ( x  t )

Thus, we have proved the following theorem:

Theorem:

The presence of the noise f (x ) in the input function f (x ) perturbs the solution vector y (x ) by an
amount equivalent to the solution of the Abel integral equation (16) with input equal to the noise
f (x ) itself.
274 Sunil Kumar et al.

As f (x ) is not known before hand, we take an upper bound for f (x ). Let sup f ( x )   , then
0  x 1
x y (t )
(16) reduces to cy ( x)      dt , which can be readily solved.
0 ( x  t )

5. Illustrative Examples
In this section, we discuss the implementation of our proposed algorithm and investigate its
accuracy by applying the homotopy perturbation method. The simplicity and accuracy of the
proposed method is illustrated through the following numerical examples by computing the
absolute error E i ( x )  y i ( x )  ~
y i m ( x ) , 1  i  n , where y i (x ) is the exact solution and
~
y i m ( x ) is the approximate solution of the problem when the series (13) is truncated at level
j  m.

Example 1.

As the first examples, we consider the following system of generalized Abel’s integral equations

 x 3
 x 4x 2
y1(x)  y2 (x)  
1
 y1(t)  y2 (t) dt  x  x   ,
 x  t 2 3

0
(17)

x
x
2 y ( x)  y ( x)  
1
 y (t ) dt  2 x  x  ,

2 1 2
 0 x t 2

with the exact solutions y1 ( x )  x , and y 2 ( x )  x . This is a weakly singular system of

Volterra integral equations of convolution type.
3
4x 2 x x
Choosing the initial approximations L10 ( x)  x   , L2 0 ( x)  x  and using the
2 3 4
iterative scheme (12), we obtain the various iterates as follows:

3
 x 4x 2 3  x2
p : L11( x)  
1
  x 2  ,
2 3 2
3
x x 2
L21( x)    ,
4 6
AAM: Intern. J., Vol. 6, Issue 1 (June 2011) [Previously, Vol. 6, Issue 11, pp. 2009 – 2024] 275

5
7 2 x2 8 x 2
3  x2
p : L12 ( x)   x 
2
 2
 ,
2 16 15
3
x 2
 2 x2
L2 2 ( x)   ,
6 32
5 5
7 2 x2 8 x 2  2 x 2  2 x3
p : L1 3 ( x)  
3
   ,... .
16 15 2 6
5
2 x 2
 2 x2
L2 3 ( x)    ,... .
60 32

Figures 1 and 2 show the absolute error between the exact solution y i ( x ) and the approximate
solution ~
yi 28 (x) obtained by truncating (13) at level n  28 for i  1, 2, respectively, for
example.

8. 10 6

6. 10 6

4. 10 6

2. 10 6

0.2 0.4 0.6 0.8 1.0

Figure 1. The absolute error E1 ( x) for Example 1

1.2 10 14

1. 10 14

8. 10 15

6. 10 15

4. 10 15

2. 10 15

0.2 0.4 0.6 0.8 1.0

Figure 2. The absolute error E2 ( x) for Example 1

276 Sunil Kumar et al.

Example 2.

In this example, the following system of generalized Abel’s integral equations of the second kind
is considered.
 x
2 y1 (t ) x
y2 (t) 1 2 x
 12 y ( x)  y 2 ( x)  0 x  t dt  0 1
dt   4 x  x 3
 ,
 (x  t) 3 3 3


13  3  4
x 12     (18)

x
y2 (t) 1  4  3 ,
 y2 ( x)  y 1 ( x)  0 (x  t) 14 dt  x 3
 ex erfc ( x )  1
 25
  
  12 

x 1
with the exact solutions y1 ( x ) 1  e erfc ( x ) and y 2 ( x )  x 3
.

13 3 4
x 12
    
x  4   3  , the iteration formula (12) gives
Taking, L10 ( x)  2 x 
1
and L2 0 ( x)  x  3

3 3  25 
 
 12 
the various Lij (x) as

3 7
x
4 x 2 16 x 4
p : L11( x)   x 
1
  ,
3 3 9 3 63 3
 3  3 
 x 4  
13  3  4 1  4  ,
L21( x)  x 12      
 4   3    25   17  
  12    
6 

3 9  11
3 3 7 5 2 16 2 x 4  
4 x 2 4 x 2 16 x 4  x 8x 2    3  2 2
 4 ,
p2 : L12 ( x)          
3 9 3 63 3 6 3 45 3   4    13
63 3  
4
 3  3 
4
x  
 4   1
2
11   3   4  ,
L22 ( x)  x 6       
  4   3    17   43 
  6    
 12  
AAM: Intern. J., Vol. 6, Issue 1 (June 2011) [Previously, Vol. 6, Issue 11, pp. 2009 – 2024] 277

5 11 5 2 3 2
 2 x2  2 x2 8 2 x
64 2 x 4 8x 2    3    2 x3   3  
2
p : L13 ( x)  
3
            
2 6 3 45 3 693 3 45 3   4   18 3   4  
3
13   3  
3 119
16 2 x 4   x 4    
  4   4  ,
 13   17 
63 3  3 3  
4 4

 3  3 
3
 x 4   
31   3   4 1  4   ,... .
L23 ( x)  x 12
       
  4   3    43  13  
  12    
3 

Figures 3 and 4 show the absolute errors between the exact solution y i ( x ) and the approximate
solution ~
yi 30 (x) for i  1, 2, respectively, for Example 2.

0.000014

0.000012

0.00001

8. 10 6

6. 10 6

4. 10 6

2. 10 6

0.2 0.4 0.6 0.8 1.0

Figure 3. The absolute error E1 ( x) for Example 2

2.5 10 16

2. 10 16

1.5 10 16

1. 10 16

5. 10 17

0.2 0.4 0.6 0.8 1.0

Figure 4. The absolute error E2 ( x) for Example 2
278 Sunil Kumar et al.

Example 3.

In this example, we consider the following system of generalized Abel’s integral equations

 x x 12 8
49x 7 175x 5  13 9 x4 
1F2 1, 
7 y1(t) 5y2 (t)
35y1(x)  y2 (x)   dt   dt 35sinx  x   , , ,
  10 5 4 
2 2
0 (x  t)
5
0 (x  t)
7 12 24
 (19)
 y2 (t) dt 5x sinx 25x ,
9
 x 5

 5y2 (x)  y1(x)   1

 0 (x  t)
5 36

with the exact solutions y1(x)  sinx and y2 (x)  x.

What follows is self explanatory.

12 8
7 x 7 5x 5  13 9  x4 
p : L10 ( x)  sin x   1F2 1,  , , ,
0

60 24  10 5  4 
9
5x 5
L20 ( x)  x  ,
36

81  3   19  88  5   14 
12 7x    5x 35   
35 8
 7   5   5 x 5 F 1, 13 , 9 , x 
4
p1 : L11 ( x)  
7x

7
5  7  
1 2   
60  116  123 24  10 5  4 
300   252  
 35   35 
2
  3 11
x     5

  5   8 21 x 4 
 1 2 1,  ,
F , ,
 16   5 10 4 
25  
5
2
  4 
13
9 x    
5

L21 ( x)  
5x 5
   5   ,... .
36  18 
25  
5

Figures 5 and 6 show the absolute errors between the exact solution y i ( x ) and the approximate
solution ~
yi10 (x) obtained by truncating (13) at level n  10 for i  1, 2, respectively, for
Example 3.
AAM: Intern. J., Vol. 6, Issue 1 (June 2011) [Previously, Vol. 6, Issue 11, pp. 2009 – 2024] 279

3.5 10 7

3. 10 7

2.5 10 7

2. 10 7

1.5 10 7

1. 10 7

5. 10 8

0.2 0.4 0.6 0.8 1.0

Figure 5. The absolute error E1 ( x) for Example 3

0.00001

8. 10 6

6. 10 6

4. 10 6

2. 10 6

0.2 0.4 0.6 0.8 1.0

Figure 6. The absolute error E2 ( x) for Example 3

Example 4.

Now we consider the following system of generalized Abel’s integral equations



 14
2 5x 50 x 125 x
5 x y1(t) x 25y (t) x 4y (t) 1 5
100y1(x)  y2(x)  y3(x)   dt   2
dt  
2
dt 100 x  x 3  x    ,
  0 (x  t) 1
0 ( x  t) 3
1
0 (x  t) 5 2 3 3 63

  3   4  1312
 8 9     x (20)
x 9y (t) x 4y (t) 1
2 27 x
3
 4  3
 36y2(x)  y1(x)  y3(x)   dt    dt  36 x  x  x  
2 3 3 ,

1
0 (x  t) 4
1
0 (x  t) 3 10  25
 
  12

  5 4
20 11    x 3
 x 11y1(t) x 4y (t) 1 343 x 7
 6
44 y3(x)  y1(x)  y2(x)   dt   3 1 dt  44x2  x  x 3   ,

1
0 (x  t) 6 0 (x  t ) 7 195  7
2  
  3

1
with the exact solutions y1(x)  x, y2 (x)  x 3 and y3 (x)  x2.
280 Sunil Kumar et al.

14
x  x 5x 5
p 0
: L1 0 ( x )  x    ,
40 6 3 252
13 3  4
8 x 12
   
L20 (x)  x 
1
3 3x 3
 4  3
,
40  25 
4 
 12 
4 5 
20  x 3 
L30 ( x)  x 2 
343 x 7
 6 ,
8580 7
8  
3

10  2   4   5  3215
3 3  x 3
3       x
7 14
x x x x 2 x 4 5x
2 2
 3  5  6 5
p : L11(x)   
1
     
40 6 3 90 3 600 63 3 252 1 7  47
560 2 3   200  
 6  15
33 19  4   27 128
x 10  343   x 35
 5   5  7  ,
 43 163
1008    214500  
 10  35 
2
 3   4  x116   3    4 
13 41
12   
3 11
8 x       
12
   3x    
3x 3
 x  2 5
2
 4  3    4   3   4  3 
L21 ( x)       
40 144  3   6   25   17   53 
4  16  160 
 12  6  12 
74  2   27 
343x 21   
  3  7 ,
 95 
77220 
 21

4 5 46 5 6 109  5   19 

11 11 20  x 3    x 21   5 x 30    
9x 3 x
6 6
343 x 7
6  6 7  6  5 
L31 ( x )     
2200 110 8580 7  67   139 
8  88  1008  
3  21   30 
2
26   6 
2 x    
7

   7   ,.... .
 33 
121  
 7 
AAM: Intern. J., Vol. 6, Issue 1 (June 2011) [Previously, Vol. 6, Issue 11, pp. 2009 – 2024] 281

Figures 7 and 9 show the absolute errors between the exact solution y i ( x ) and the approximate
solution ~
yi 7 (x) obtained by truncating (13) at level n  7 for i  1, 2, 3, respectively, for
Example 4.

4. 10 6

3. 10 6

2. 10 6

1. 10 6

0.2 0.4 0.6 0.8 1.0

Fig.7. The absolute error E1 ( x) for Ex.4.
4. 10 6

3. 10 6

2. 10 6

1. 10 6

0.2 0.4 0.6 0.8 1.0

Figure 8. The absolute error E2 ( x) for Example 4

3. 10 6

2.5 10 6

2. 10 6

1.5 10 6

1. 10 6

5. 10 7

0.2 0.4 0.6 0.8 1.0

Figure 9. The absolute error E3 ( x) for Example 4
282 Sunil Kumar et al.

5. Conclusion

The Homotopy perturbation method proposed is a new and efficient algorithm for the numerical
solution of system of generalized Abel’s integral equations. It is proved that the change
 y ( x ), in the solution y ( x ) caused by the presence of noise  f ( x ) in the observable data
f ( x ), is the solution of the generalized Abel integral equation with input data equal to the
noise  f ( x ) itself. From the given numerical examples, and Figs 1-9, we conclude that the
method is accurate and easy to implement for solving systems of generalized Abel’s integral
equations especially of the second kind.

Acknowledgment

The first author acknowledges the financial supports from Rajiv Gandhi National Fellowship of
the University Grant Commission, New Delhi, India under the JRF schemes.

REFERENCES

Abbasbandy, S. (2006). Numerical Solution of the integral equations: Homotopy Perturbation

Method and Adomian Decomposition method. Applied Mathematics and
Computation, 173, pp. 493-500.
Abbasbandy, S. (2007).Application of He’s Homotopy Perturbation Method to Functional
Integral Equations. Chaos, Solitons & Fractals, 31 pp. 1243-1247.
Hossein, Aminikhah and Salahi, Maziar (2009). A New HPM for Integral Equations, Appl.
Math., 4(1) pp. 122-133.
Biazar, J., Eslami, M. and Aminikhah, H. (2009) Application of Homotopy Perturbation
Method for Volterra Integral Equation of first kind, Chaos, Solitons & Fractals 42 pp.
3020-3026.
Biazar, J. and Eslami, M. (2010). An Efficient Technique for Solving Special Integral Equations,
Appl. Appl. Math., 5(1) pp. 217-224.
Burton, T. A. (2005) .Volterra integral and differential equations, Elsevier B. V., Netherlands.
Dehghan, Mehdi and Shakeri, F. (2008). Approximate solution of a differential equation arising
in astrophysics using the variational iteration method, New Astronomy 13, pp. 53-59.
Dehghan, Mehdi and Shakeri, F. (2008). Solution of an Integro- Differential equation arising in
oscillating Magnetic fields using He’s Homotopy Perturbation Method, PIER 78, pp. 361-
376.
Ganji, D.D. and Rajabi, A. (2006). Assessment of Homotopy Perturbation and Perturbation
Methods in heat radiation equations, Int. Commun. Heat Mass Transfer 33, pp. 391- 400.
Ganji, D.D., Afrouzi, G.A., Hossenzadeh, H. and Talarposhti, R.A. (2007). Applications of
Homotopy Perturbation and Perturbation to the second kind of nonlinear integral equations,
Phys. Lett. A. 371, pp. 20-25.
AAM: Intern. J., Vol. 6, Issue 1 (June 2011) [Previously, Vol. 6, Issue 11, pp. 2009 – 2024] 283

He, J.H. (1998). An approximate solution technique depending upon an artificial parameter,
Commun. Nonlinear Sci. Simul. 3(2), pp. 92-97.
He, J.H. (1999). Homotopy perturbation technique, Comput. Methods Appl. Mech. Eng. 178, pp.
257-262.
He, J.H. (2000). A coupling method of homotopy technique and perturbation technique for
nonlinear problems, Int. J. Nonlinear Mech. 35, pp. 37-43.
He, J.H. (2000). A review on some new recently developed nonlinear analytical technique, Int. J.
Nonlinear Sci. Numer. Simul. 1(1), pp. 51-70.
He, J.H. (2004). The homotopy perturbation method for nonlinear oscillator with discontinuities,
Appl. Math. Comput. 151, pp. 287-292.
Lalescu, T. (1912). Introduction a la theories des equations integrals. Avec une preface de E.
Picard, Paris: A. Hermann et Fils,. VII + 152 pp.
Lio, S. J. (1995). An approximate solution technique not depending on small parameters: A
special examples, Int. J. Nonlinear Mech. 30(3), pp. 371-380.
Lio, S. J. (1997). Boundary element method for general nonlinear differential operators, Eng.
Anal. Boundary Elements 20, pp. 91-98.
Maleknejad, K., Kajani, M. T. and Mahmoudi, Y. (2003). Numerical solution of Fredholm and
Volterra integral equation of the second kind by using Legendre Wavelets, Kybernetes 32
(9-10), pp. 1530-1539.
Maleknejad, K. and Kajani, M. T. (2004). Solving linear integro-differential equation system by
Galerkin methods with hybrid functions, Appl. Math. Comput. 159, pp. 603-612.
Maleknejad, K. and Aghazadeh, N. (2005). Numerical solution of Volterra integral equations of
the second kind with convolution type kernel by using Taylors- series expansion method,
Appl. Math. Comput. 161, pp. 915-922.
Maleknejad, K. and Sohrabi, Rostami, S. Y. (2007). Numerical solution of nonlinear Volterra
integral equation of the second kind by using Chebyshev polynomials, Appl. Math. Comput.
188, pp. 123-128.
Pandey, R. K., Singh, O. P. and Singh, K. (2009). Numerical solution of system of Volterra
integral equations using Bernstein polynomials, Nonl. Sci. and Simul. 10(7), pp. 891-895.
Rabbani, M. K., Maleknejad and Aghazadeh, N. (2007). Numerical computational solution of
the Volterra integral equations system of the second kind by using an expansion method,
Appl. Math. Comput.187, pp. 1143-1146.
Shakeri, F. and Dehghan Mehdi, (2008). Solution of Deley differential equation via a Homotopy
perturbation method, Mathematical and computer Modelling, 48, pp.486-498.
Singh, K., Singh, O. P. and Pandey, R. K. (2010). Almost Bernstein operational matrix method
for system of Volterra integral equations of convolution type, Nonl. Sci. Lett, A Vol. 1 No.
2, pp. 201-206.
Tahmasbi, A. and Fard, O. S. (2008). Numerical solution of linear Volterra integral equations
system of the second kind, , Appl. Math. Comput. 201, pp. 547-552.
Yalsinbas, S. (2002). Taylor polynomials solutions of nonlinear Volterra–Fredholm integral
equations, Appl. Math. Comput. 127, pp. 195-206.