Professional Documents
Culture Documents
Var Alg
Var Alg
Var Alg
2012
Variational iteration method is widely used for solving nonlinear, integro-differential equations. Through
careful investigation of the iteration formulas of the earlier variational iteration algorithm (VIM), we find
unnecessary repeated calculations in each iteration. To overcome this shortcoming, the variational
iteration algorithm-II (He et al., 2010) will be used in this paper to solve the system of non linear integro-
differential equations. Beside this, the comparison of the exact solution with approximated solution by
VIM-II is illustrated by the graphs. Several examples are given to verify the reliability and efficiency of
the method.
Key words: Variational iteration algorithm-II (VIM), system of nonlinear integro-differential equations.
INTRODUCTION
The higher-order integro differential equations arise in integro-differential equations using the variational
mathematical, applied and engineering sciences, iteration method. The approach used is based on the
astrophysics, solid state physics, astronomy, fluid variational iteration method (VIM) proposed by the
dynamics, beam theory, fiber optics, glass-forming Chinese researcher J. H. He. This method has wider
process and chemical reaction-diffusion models application because; it reduces the size of computation
(Agarwal, 1983; Hashim and Comput, 2006; Morchalo, and is a very powerful mathematical tool for various kinds
1975; Kythe and Puri, 2002; Wang et al., 2007). Several of linear and nonlinear problems as well as system of
techniques including Adomian decomposition, variational ordinary differential equations, partial differential
iteration and homotopy perturbation method have been equations and integro-differential equations. In a recent
used to investigate integro-differential (Hashim and publication, He et al. (2010) stated that “The Variational
Comput, 2006; El-Sayed et al., 2004; Saberi-Nadjafi and Iteration Method Which Should be Followed”, the problem
Tamamgar, 2008; Abbasbandy and Shivanian, 2009; was completely eliminated, and a new iteration algorithm
Abbasbandy et al., 2008, 2009; Sweilam, 2007; Xu et al., was suggested, and the algorithm was termed as the
2009; Mohyud-Din and Naturforsch, 2010; Yildirim and variational iteration algorithm-II. To the best of our
Naturforsch, 2010; Wang and He, 2007; Hesameddini knowledge, this paper represents the first application of
and Latifizadeh, 2009; Ghasemi et al., 2007; Yusufoglu, variational iteration algorithm-II to solve the system of
2007; Yildirim, 2008; El-Shahed, 2005). nonlinear integro-differential equations of third order.
In a recent review article, He et al. (2010) summarized Compared with the classical iteration algorithm, given by
three variational iteration algorithms. Here, we will use Saberi-Nadjafi and Tamamgar (2008), Abbasbandy and
the variational iteration algorithm-II for the study. It is an Shivanian (2009) and Abbasbandy et al. (2008, 2009),
alternative approach to system of third order non-linear the present one is much more concise and effective and
no unnecessary repeated calculation is needed as shown
in the given examples. The elegance of variational
iteration algorithm-II can be attributed to its simplistic
*Corresponding author. E-mail: yasirmath@yahoo.com. approach in seeking the analytical solution of the system
Khan et al. 255
of nonlinear integro-differential equations, given as: correction functional (3) stationary, and noticing
~
that f n 0 ,
f1m x G1 x, f 2 x,..., f 2m x, f 3 x,..., f 3m x,..., f n x,..., f nm x
x
0 0
x
f i ,k 1 x m f i ,k x mi s [ f i ,km s G
m n ( s, f1,k x ,..., f1,k s ,..., f i 1,k s ,..., f i 1,k s , f i 1,k s ,..., f i 1,k s
~ ~ m ~ ~ m ~ ~ m
f 2 x G2 x, f1 x,..., f1 x, f 3 x,..., f03 x,..., f n x,..., f n x
m
x
H 2 x, t , f1 t ,..., f1m t ,..., f n t ,..., f nm t dt
s
,..., f n,k s ,..., f n,mk s ) H n s, t , f1 t ,..., f1m t ,..., f n t ,..., f nm
(1) ~ ~ ~ ~ ~ ~
0
0
f nm x Gn x, f1 x,..., f1m x,...,
, f 2 ~fx,...,sf2,...,
xf,..., fns1)x,..., fHnm1 xs, t , ~f t ,..., ~f m t ,..., ~f t ,..., ~f m t dt ]ds 0
s
m ~ m
n,k n,k n 1 1 n n
x
H n x, t , f1 t ,..., f1m t ,..., f n t ,..., f nm t dt
0
or
0
x
f i ,k 1 x f i ,k x i s f i ,km s ds 0
with initial conditions f i
k
bi cki , k 0,1,2,..., m 1,
0
i 1,2,..., n. (2)
or
In system (1), f i x is the solution to be determined, m
f i ,k 1 x f i ,k x i s f i ,km1 s s x 11 i1 s
is the order of derivatives and Gi , H i i 1,2,..., n are
f i ,k 1 x functions.
known linear ornonlinear f i ,k x i s f i ,km1 s 1 i1 s f i ,km2 s
1
sx sx
x
1 i
m 1 m 1
s fi,k s s x 1 i m s fi,k s ds 0
m
According to VIM, Equations [7 to 10], we can construct a for i 1,2,3,..., n , yields the following stationary
correction functional of system (1) as: conditions:
~ f i,k s ~ m : i ~ s 0, ~ m
x m
f i ,k 1 x f i ,k x i s [ f i ,k s Gn (s, f1,k x,..., f1,k s ,..., f i1,k s ,..., f i1,k s , f i1,k s ,..., f i1,k s
m ~ ~ m
1
m x
~ m Where ~ 1 s ~and ~2 s are ~Lagrange multiplier,
f i ,k 1 x f i ,0 x
~ m m
m 1! 0
m 1
s x [G ( s , f x ,..., f s ,..., f s ,..., f s , f s ,..., f
~s are restricted variations. Imposing s
and f s and g
n 1, k 1, k ~i 1, k i 1, k i 1, k i 1, k
1 s 2 s s x 2
n 1 1 n n (4) 1
(9)
for i 1,2,3,..., n 2
In order to elucidate the solution procedure of the As a result, we obtain the following iteration formula:
variational iteration algorithm-II, we consider the system
of third order non-linear integro-differential equations.
1 2
x
f n 1 x f n x s x f n s e s s g (
1 1
0
Example 1. Consider the following 3rd order non-linear 2 2 2
integro-differential system.
x s 1 2
x
g 0 x 1 x 2
1
f 0 x 1 2 x x 2
The exact solution of the problem is: 1
and (11)
f x x e x , g x x e x (7) 2 2
According to the classical variational iteration method, This satisfies the initial conditions in Equation 6.
(Saberi-Nadjafi and Tamamgar 2008; Abbasbandy and By the iteration formula (Equation 10) and the initial
Shivanian, 2009; Abbasbandy et al., 2008; Sweilam, solution (Equation 11), we have the following first order
2007), the correction functional can be written in the approximation solutions
following form
x 2 f s e s s
x
x
1 1
s
2 f 1 x1 f0 2 x
f n1 x f n x 1 s f (n ) s e s g n s f n t g n t dt ds 0
s 2 s
1
1
0
2 2
2 2 2s 0
x
0 1 1 1
f n11 x 1f n x x11 ss x f (n ) s e s s g n 2 s s fn 2 t g n 2 t s dt ds
n ) s e s g n 2 s 0 s fn 2 1t g n 2s22t dtds 1 ds 2
2 1 1 2 02 1 22
s
1 g s 0 x0
g1 n21 xf 12gx n xf 1x22 s g ( n ) ss xe 1
f f n s s
t g t dt 0ds
e s
f
n t g
s g s
n t dt f t g 2
t dt ds
g2 n1 x 2 g n x1s 220 s 0 g (n2) s e s 1 fn s 2fn 2 t 2g n 2 t dt2 ds
s
2 0 2 2 4s 0 ( 0)
( 0) ( 0)
s e s f
) n n n 1
0
( n ) s e 1 f n s s 0f n t g n t dt ds
s 2
2 4 0
x4 x5 x6
s
40
14 0 2
n ) s e 1 f n s f n t g n t dt ds 2
e x
x
24 60 360
(12)
(8)
Khan et al. 257
2 1 f x 1 2x 1 x2
g 1 x g 0 x s x g (0 ) s 1 f 0 s 0 f (02) t g2 (02) t dt ds
1
x s
0
2 40 1 2
g0 x 1 x
s 2
g (0 ) s 1 f 0 s f (02) t g (02) t dt ds
1
x 4 x5 x6
40 f1 x = e x x
4 5 24 60 360
x x
e x x (13) g1 x e x x
x 4 x5
24 120 24 120
7x 2 x3 17e x x3 7e x x 4 e x x5
Continuing in this manner, we obtain the other f 2 x 81 80e x 29x 53e x x 16e x x 2
components of f n1 x and g n1 x
2 6 6 24 60
x7 x8 x9 x10 x11 41x12 x13 x14
5040 5760 45360 181440 380160 68428800 12355200 172972800
7x 2 x3 17ex x3 7e x x 4 e x x5
f 2 x 81 80e x 29x 53e x x 16e x x 2 47 49e x 19x 5x 2 9e x x 2 x3 5e x x3 5e x x 4 x5 e x x5
2 6 6 24 60 g 2 x 15e x x
7 8 9 10 11 12 13
2 2 2 4 2 12 6 48 120 120
x x x x x 41x x x14
x6 x7 x8 x9 x11 31x12 x13 x14
5040 5760 45360 181440 380160 68428800 12355200 172972800
360 2520 80640 362880 2280960 136857600 24710400 345945600
47 49e x 19x 5x 2 9ex x 2 x3 5e x x3 5e x x 4 x5 ex x5
g 2 x 15ex x 233531999983989 28507321371ex 23329141 e2x 16657555191939x
2 2 2 4 2 12 6 48 120 120 f 3 x
16384 2 16384 8192
x6 x7 x8 x9 x11 31x12 x13 x14
2x 2
6629497e x 831700590567x 10389928139e x x 2
360 2520 80640 362880 2280960 136857600 24710400 345945600 12220268945e x x ...
4096 8192 2
233531999983989 28507321371ex 23329141 e2x 16657555191939x
f 3 x 157150177236793 19183368817e x 19855161e2x 11312112683627x
16384 2 16384 8192 g 3 x
2x 2 x 2 32768 4 32768 16384
6629497 e x 831700590567 x 10389928139 e x
12220268945e x x ... x 2x
16421624645e x 5650023 e x 570608576875x 2
4096 8192 2 1742312038e x x 2 ...
4 8192 16384
157150177236793 19183368817e x 19855161e2x 11312112683627x
g 3 x (16)
32768 4 32768 16384
16421624645e x x 5650023 e2x x 570608576875x 2 and so on. In a similar manner, the rest of the
1742312038e x x 2 ...
4 8192 16384 components can be obtained by using Equation 15.
(14)
Example 2. Consider the following 3rd order non-linear
According to VIM-II (He et al., 2010; He and Wu, 2007), integro-differential system.
the iteration formula for Equation (5) is
x
f 1 x 12 f 2 x g x x t g t f 2 t dt
x 1
1 2
x s
1 1
f n1 x f 0 x s x e s gn s f n t2 2gn t dt ds0
s 2
0x x
2 2 2 20
2
s
ds
x t f t g 2 t dt
1 1 1 1 x 1
f 1n1 x 2 f 0 1 x x 2 1 s x 22 e s s gn 1 ss g 2x fn t2 f g xn gt x dt
s s 2 2 2
1
e s g g nx s g x f n 2 t sgx n t e dt12ds
s
2
n1 0 2 0 0 f n 2s f n2 0 t 2gn 2 t dt ds 0
2
2 40 (17)
2 2 s
s 0x
2 1
s
1 1 1
s g g x1 s g x f t s g x t edt 1ds f n s 1
e s 21 f n 2 ns1 n f 02n 20 t ng2n 2 t n dt ds
f n t gn t dt ds
s 2 s 2 2
4 with initial conditions
40
0 0
with the starting of the initial approximation The exact solution of the problem is
Using Equation 4 and by selecting f 0 x x and u0 1 , u0 0 , u0 1 and v0 0 , v0 1 ,
v0 0
g 0 x 1
1 2
x , we obtain
2
The exact solution is
2
1
x s
s 1
f 1 x x s x 1 f0 s g0 s x t g0 t f02 t dt ds ux Cosx , vx Sinx (23)
0 2 2
2 0
3 5 6 7
x x x x Its iteration formulation is suggested according to He et al.
= x
6 120 480 5040 (2010), and He and Wu (2007).
2 s 1
x2 1
x s
g 1 x 1 s x f0 s g0 s x t f0 t g02 t dt ds 2
1
x s
2 02 2 2 0 un1 x u0 x s x s vn s un2 t vn2 t dt ds
0
2 4 6
2 0
x x x
= 1
2
1
x s
2 24 1440 vn1 x v0 x s x s 2 un s s vn2 t u n t vn t dt ds
0
(20) 2 0
(24)
Similarly, one can obtain the second-order, third-order
and fourth-order approximate solutions given below
Beginning with
3 5 7 9 10 11 12
x x x x x 83x x
f 2 x x x2
6 120 5040 290304 2073600 79833600 19958400 u 0 x 1 and v0 x x (25)
7x 13
157x14 x15 x16 2
197683200 19372953600 1257984000 42268262400
We have
x2 x4 x6 x8 x9 x10 x11 47x12
g 2 x 1
2 24 720 40320 483840 691200 1774080 479001600
2
1
x s
164055193x16 226161763x18 42090421x 20
u x x v x u2 t v2 t dt
27172454400 140107968000 79394515200
...
0
x3 7x5 x 7 667x9 443x11 101213x13 322403x15
x (22) v3 x x
2
v x x u x x v t u t v t dt
2
6 120 1680 51840 1330560 16174080 63866880
584431x17 7415129x19
0 ...
164975616 21555072000
with initial conditions (26)
Khan et al. 259