See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/259143453

Solving BVPs with shooting method and VIMHP

Article  in  Journal of the Egyptian Mathematical Society · October 2013

DOI: 10.1016/j.joems.2013.04.005

CITATIONS READS

4 3,637

2 authors:

Mashallah Matinfar Maryam Ghasemi

University of Mazandaran University of Waterloo
94 PUBLICATIONS   491 CITATIONS    20 PUBLICATIONS   71 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Mesh less method, Numerical method View project

All content following this page was uploaded by Mashallah Matinfar on 04 October 2016.

The user has requested enhancement of the downloaded file.

Journal of the Egyptian Mathematical Society (2013) 21, 354–360

Egyptian Mathematical Society

Journal of the Egyptian Mathematical Society

www.etms-eg.org
www.elsevier.com/locate/joems

ORIGINAL ARTICLE

Solving BVPs with shooting method and VIMHP

M. Matinfar *, M. Ghasemi

Science of Mathematics Faculty, Department of Mathematics, University of Mazandaran, PO Box 47415-95447, Babolsar, Iran

Received 30 January 2013; revised 8 April 2013; accepted 15 April 2013

Available online 17 May 2013

KEYWORDS Abstract In this paper, a new method is applied for solving the nonlinear Boundary value prob-
Nonlinear boundary value lems. This method is a combination of shooting method and Variational Iteration Method Using
problems He’s Polynomials. As examples show, our proposed technique can overcome the difficulties that
Shooting method arise in both methods, and efficiency of this technique is approved.
Variational iteration method
He’s polynomials 2010 MATHEMATICS SUBJECT CLASSIFICATION: 65L10, 65N99, 65L99

ª 2013 Production and hosting by Elsevier B.V. on behalf of Egyptian Mathematical Society.
Open access under CC BY-NC-ND license.

1. Introduction and then using numerical methods such as Runge-Kutta, the

Solving nonlinear differential equations are very important be-
cause they have the ability to model most phenomena in the
world. So, scientist and researchers are interested finding the
best way to determine the solution of nonlinear ODEs and
PDEs. One group of differential equations is Boundary value
problems [1] that can be solved by numerical methods. As
we know, besides the ability of numerical techniques to find
the solution of differential equations, they need huge computa-
tional work and are very time consuming. One numerical tech-
nique is shooting method that is used to solve BVPs. In this
method by choosing the arbitrary value for derivatives of de-
sired function in staring time and converting BVPs into IVPs,
* Corresponding author. Tel.: +98 1125342430.
E-mail address: m.matinfar@umz.ac.ir (M. Matinfar).
Peer review under responsibility of Egyptian Mathematical Society.
boundary value problem can be solved. Unfortunately, this
procedure is not easy to apply. Therefore, it seems using ana-
lytical method can overcome this problem. But we encounter a
problem in most of the analytical methods like Variational
Iteration Method (VIM) [2–4], Homotopy perturbation Meth-
od (HPM) [5–7] and Adomian Decomposition Method
(ADM) [8–10] for solving BVPs. In these methods, in order
to get better results, we should choose the appropriate initial
guess to start the recursive procedure, and this is difficult for
BVPs. In this paper, we combine the VIMHP as a convergent
and powerful method with shooting method to solve BVPs.
This modification overcomes the mentioned difficulties that ex-
ist in numerical and analytical techniques for solving BVPs.
The accuracy of the proposed method is approved via solving
some examples and comparing the obtained results with the
solutions of other methods.
2. Methodology

In following, the concepts of VIMHP, shooting method and

Production and hosting by Elsevier our proposed method are presented:
1110-256X ª 2013 Production and hosting by Elsevier B.V. on behalf of Egyptian Mathematical Society. Open access under CC BY-NC-ND license.
http://dx.doi.org/10.1016/j.joems.2013.04.005
Solving BVPs with shooting method and VIMHP 355

2.1. Variational iteration method the domain X. Generally speaking the operator A can be di-
vided into two parts L and N, where L is a linear, and N is a
To illustrate the basic idea of VIM, at first consider the follow- nonlinear operator. Therefore (3) can be rewritten as follows:
ing nonlinear differential equation LðUÞ þ NðUÞ  fðrÞ ¼ 0:
L½uðrÞ þ N½uðrÞ ¼ gðrÞ; r > 0; ð1Þ
We construct a homotopy V(r, p): X · [0, 1] fi Rn, which
dm satisfies
where L ¼ ; m 2 N, is a linear operator as, N is a nonlinear
drm
operator and g(r) is the source inhomogeneous term, subject to HðV; pÞ ¼ ð1  pÞ½LðVÞ  LðU0 Þ þ p½AðVÞ  fðrÞ ¼ 0; p
the initial conditions
2 ½0; 1; r 2 X;
uðkÞ ð0Þ ¼ ck ; k ¼ 0; 1; 2; . . . ; m  1; ð2Þ
where ck is a real number. According to the He’s variational or equivalently,
iteration method [11], we can construct a correction functional HðV; pÞ ¼ LðVÞ  LðU0 Þ þ pLðU0 Þ þ p½NðVÞ  fðrÞ ¼ 0: ð4Þ
for (1) as follows:
Z r where U0 is an initial approximation of (3). In this method,
unþ1 ðrÞ ¼ un ðrÞ þ kðsÞfLun ðsÞ þ N uen ðsÞ  gðsÞgds; n P 0; using the homotopy parameter p, we have the following power
0 series presentation for V,
where k is a general Lagrangian multiplier and can be identi- V ¼ V0 þ pV1 þ p2 V2 þ . . . :
fied optimally via variational theory. Here, we apply restricted The approximate solution can be obtained by setting p = 1,
variations to nonlinear term Nu, in this case the best value of i.e.
multiplier we can be easily determined. Making the above
functional stationary, noticing that d uen ¼ 0, U ¼ U0 þ U 1 þ U 2 þ . . . : ð5Þ
Z r
The convergence of series (5) is discussed in [15]. The meth-
dunþ1 ðrÞ ¼ dun ðrÞ þ d kðsÞfLun ðsÞ  gðsÞgds; od considers the nonlinear term N[V] as
0

yields the following Lagrange multipliers, X

þ1
NðVÞ ¼ pi Hi ¼ H0 þ pH1 þ p2 H2 þ   
k ¼ 1 for m ¼ 1; i¼0
k ¼ s  r; for m ¼ 2;
where Hn’s are the so-called He’s polynomials [16], which can
and in general, be calculated by using the formula
ð1Þm !!
k¼ ðs  rÞðm1Þ ; for m P 1: 1 @n X
n
ðm  1Þ! Hn ðV0 ; V1 ;    ; Vn Þ ¼ N pi Vi ; n
n! @pn i¼0 p¼0
The successive approximations un(r), n P 0 of the solution u(r)
will be readily obtained upon using the obtained Lagrange ¼ 0; 1; 2;   
multiplier and by using any selective function u0. Conse-
quently, the exact solution may be obtained as
2.3. Variational iteration method using He’s polynomials
uðrÞ ¼ lim un ðrÞ:
n!1
To illustrate the basic idea of VIMHP, consider the following
general differential equation:
2.2. Homotopy perturbation method
L½uðrÞ þ N½uðrÞ ¼ gðrÞ; ð6Þ
We know the essential idea of HPM [12–14] is to introduce a where L is a linear operator, N a nonlinear operator and g(r) is
homotopy parameter, say p, which takes the values from 0 the source inhomogeneous term. According to VIM, for n P 0
to 1. When p = 0, the system of equations is in sufficiently sim- we can construct a correct functional as follows:
plified form, which normally admits a rather simple solution.
unþ1 ðrÞ ¼ un ðrÞ
As p gradually increases to 1, the system goes through a se- Z r  m 
quence of ‘‘deformation’’, the solution of each stage is ‘‘close’’ d
þ kðsÞ un ðsÞ þ N½ e
un ðsÞ  gðsÞ ds; ð7Þ
to that at the previous stage of ‘‘deformation’’. Eventually at 0 dsm
p = 1, the system takes the original form of equation and m
ð1Þ
where kðsÞ ¼ ðm1Þ! ðs  rÞðm1Þ . Now, applying a series of the
the final stage of ‘‘deformation’’ gives the desired solution.
power of p and then using He’s polynomials we have:
To illustrate the basic concept of HPM, consider the following
nonlinear system of differential equations Z r " ! !#
X1
dm X 1 X 1
n n n
AðUÞ ¼ fðrÞ; r 2 X; ð3Þ p vn ¼ u0 þ p kðsÞ m p vn ðsÞ þ N p vn ðsÞ ds
n¼0 0 dr n¼0 n¼0
with boundary conditions Z r Z r " !
  dm X1
n
@U  kðsÞgðsÞds ¼ u0 þ p kðsÞ m p vn ðsÞ
B U; ¼ 0; r 2 C; 0 0 dr n¼0
@n # Z r
X 1
where A is a differential operator, B is a boundary operator, þ pn Hn ds  kðsÞgðsÞds; ð8Þ
0
f(r) is a known analytic function, and C is the boundary of n¼0
356 M. Matinfar, M. Ghasemi
 
which is the modified variational iteration method using He’s d2 y dy
polynomials [17–19]. Now, equating coefficients of like powers ¼ f t; y; ; ð15Þ
dt2 dt
of p, we have
on t 2 [a, b] with the general boundary conditions
Z r
p0 : v0 ¼ u0  kðsÞgðsÞds; yðaÞ ¼ a; ð16Þ
0
Z r  m  yðbÞ ¼ b: ð17Þ
ð1Þm d
p1 : v1 ¼ ðs  rÞðm1Þ m ðv0 ðsÞÞþ H0 ðv0 Þ ds;
0 ðm  1Þ! dr The methods considered thus far for second-order differential
Z r m  m 
ð1Þ ðm1Þ d equations involve a choice of the initial conditions y(a) and
p2 : v2 ¼ ðs  rÞ m ðv1 ðsÞÞþ H1 ðv0 ; v1 Þ ds;
0 ðm  1Þ! dr y0 (a). We can still approach the boundary value problem from
.. this framework by choosing the ‘‘initial’’ conditions
.
Z r  m 
ð1Þm d yðaÞ ¼ a; ð18Þ
j
p : vj ¼ ðs  rÞðm1Þ m ðvj1 ðsÞÞþ Hj1 ðv0 ; v1 ;    ; vj1 Þ ds;
0 ðm  1Þ! dr dyðaÞ
.. ¼ A; ð19Þ
. dt
ð9Þ where the constant A is chosen so that as we advance the solu-
Therefore, the approximated solutions of (6) can be ob- tion to t = b, y(b) = b can be found. The shooting method
tained as follows: gives an iterative procedure with which we can determine this
constant A. Fig. 1 illustrates the solution of the boundary va-
X
1
lue problem given two distinct value of A. In this case, the va-
u ¼ lim pn vn ¼ v0 þ v1 þ v2 þ . . . : ð10Þ
p!1
n¼0
lue of A = A1 gives a value for the initial slope which is too
low to satisfy the boundary conditions, whereas the value of
The zeroth (initial) approximation u0 can be freely chosen if it A = A2 is too large to satisfy (17) and both of them should
satisfies the initial and boundary conditions of the problem. be corrected.
The success of the method depends on the proper selection
of the initial approximation u0. However, using the initial val-
Theorem 2.1. The boundary value problem
ues u(k)(0) = ck, k = 0, 1, 2, . . . , m  1 are preferably used for
the selective zeroth approximation u0. In our alternative ap- y00 ¼ fðt; y; y0 Þ; a 6 x 6 b; yðaÞ ¼ a; yðbÞ ¼ b; ð20Þ
proach we select the initial approximation u0 as:
and initial value problem
X
m1
ck k
u0 ðrÞ ¼ r: ð11Þ y00 ¼ fðt; y; y0 Þ; a 6 x 6 b; yðaÞ ¼ a; y0 ðaÞ ¼ A; ð21Þ
k¼0
k!
where A is an arbitrary constant, have unique solution on [a, b]
For more information refer to [20,21]. if
2.4. Boundary value problems: the shooting method (a) f, @f @f
and @y
@y 0 be continuous;

(b) there exist constant M such that

Many physical equations do not have specified initial condi-  
 @f 
tions, but rather some given boundary conditions. A simple   6 M;
@y0 
example of such a problem is the second-order boundary value
problem (c) there exist constant L such that
 
d2 y dy @f
2
¼ f t; y; ; ð12Þ 0< < L;
dt dt @y
on t 2 [a, b] with the general boundary conditions
dyðaÞ
a1 yðaÞ þ b1 ¼ c1 ; ð13Þ
dt
dyðbÞ
a2 yðbÞ þ b2 ¼ c2 : ð14Þ
dt
Thus the solution is defined over a specific interval and
must satisfy (13) and (14) at the end points of the interval.

2.4.1. The shooting method

The boundary value problems constructed here require infor-
mation at the present time t = a and a future time t = b. How-
ever, we need only require information about the starting time
t = a. Some efforts are then needed to reconcile the analytical
and numerical schemes with the boundary value problems pre-
sented here.
We begin by reconsidering the generic boundary value Figure 1 Solutions to the boundary value problem with y(a) = a
problem and y0 (a) = A.
Solving BVPs with shooting method and VIMHP 357

on 1 2
y00 ¼ y3 ; yð1Þ ¼  ; y0 ð1Þ ¼ A; ð24Þ
0 0 2 3
D ¼ fðx; y; y Þj a 6 x 6 b; 1 < y < 1; 1 < y < 1g:
where A is an arbitrary constant and should be determined
for the proof of this theorem refer to [22]. based on the illustrated algorithm:
1 þ 23 1
A¼ ¼ : ð25Þ
2.4.2. Computational algorithm for the combination of VIMHP 21 3
and shooting method Using this value of A, we can solve Eq. (24) by the VIMHP.
The above example that adjusting the value of A in (19) can To use the VIMHP, we should calculate the Lagrange
lead to a solution which satisfies (17). We can solve this, using multiplier:
Z t
a self-consistent algorithm to search for the appropriate value 1
of A which satisfies the original problem. The basic algorithm yðnþ1Þ ðtÞ ¼ yn ðtÞ þ kðsÞfynss  ye3 n gds ð26Þ
1 2
is as follows:
yn ¼ 0.
where y~n is considered as restricted variations, i.e. d~
1-Solve the differential equation with the initial conditions To find the optimal value of k(s), we have
y(a) = a, and y0 (a) = A with VIMHP as illustrated before. Z t  
1
The first appropriate value of A can be determined by eval- dyðnþ1Þ ðtÞ ¼ dyn ðtÞ þ d kðsÞ ynss  ye3 n ds ð27Þ
uating the slope of straight lines passing through two 1 2
points, a and b, those that lie on the function curve: or
yðbÞ  yðaÞ Z t
A¼ : dyðnþ1Þ ðtÞ ¼ dyn ðtÞ þ kðsÞdynss ds; ð28Þ
ba
1
2-Evaluates the solution yVIMHP(t) at t = b and compare
which results
this value with the target value of y(b) = b.
3-Adjust the value of A (either bigger or smaller) until a dyðnþ1Þ ðtÞ ¼ dyn ðtÞð1  k0 ðsÞÞ þ kðsÞdy0n ðsÞjs¼t
desired level of tolerance and accuracy is achieved. The Z t
Secant method for determining values of A, for instance, þ k00 ðsÞdyn ðsÞds ¼ 0: ð29Þ
1
may be appropriate:
Therefore, the stationary conditions are obtained in the fol-
yVIMHP ðb; Ak1 Þ  yðbÞ
Ak ¼ Ak1  ðAk1 lowing form
yVIMHP ðb; Ak1 Þ  yVIMHP ðb; Ak2 Þ
8 0
 Ak2 Þ; ð22Þ < 1  k ðsÞ ¼ 0js¼t ;
>
where Ak1 is the value of A in the (k  1)th step and kðsÞ ¼ 0js¼t ; ð30Þ
>
: 00
yVIMHP(b; Ak1) is the value of yVIMHP(t)Œt=b for Ak1, but k ðsÞ ¼ 0js¼t ;
as we can see in this method we need the first and second value
which results k(s) = s  t. Now we can use the VIMHP to find
of A. The first value can be obtained as illustrated in step (1)
the components of the solution.
and one choice for the second value may be 2 yðbÞyðaÞ :
ba
Z t !3
4-Once the specified accuracy has been achieved, the solu- X1
1 X1

tion is complete and is accurate to the level of the tolerance pn vn ðtÞ ¼ y0  p ðs  tÞ pn vn ds: ð31Þ
n¼0
2 1 n¼0
chosen.
The best choice of y0 is y0 ¼  23  13 ðt  1Þ and based on the
Note: VIMHP definition
Our proposed method is applicable for higher order of dif- X
1
ferential equations. It only needs the equation be transformed yðtÞ ¼ vn ðtÞ:
into a system of equations of order one and two. The equation n¼0

of order one can be solved easily by VIMHP and the equation In order to obtain the unknown yn, n = 1, 2, 3, . . . the like
of order two is solved by the VIMHP and shooting method. powers of p should be compared. So we have:
2 1
3. Applied examples p0 : v0 ðtÞ ¼ y0 ¼   ðt  1Þ;
3 3
Z
1 1 t

p : v1 ðtÞ ¼  ðs  tÞv30 ds
In this section we present some examples to show the reliability 2 1
of applied method. ðt  1Þ2 3
¼ ðt þ 7t2 þ 23t þ 49Þ;
1080
Example 1. Consider the following equations, with the Z
3 t

boundary values yð1Þ ¼  23 and y(2) = 1. p2 : v2 ðtÞ ¼  ðs  tÞv20 v1 ds

2 1
1 ðt  1Þ4 5
y00 ¼ y3 : ð23Þ ¼ ðt þ 13t4 þ 82t3 þ 338t2 þ 749t þ 737Þ;
2 466560
In this example, fðt; y; y0 Þ ¼ 12 y3 and based on the Theorem 2.1, ..
Eq. (23) has unique solution. In order to solve above equation, .
it should be converted into the initial value problem as: ð32Þ
358 M. Matinfar, M. Ghasemi

The nth approximate P of solution of Eq. (24) can be

Table 1 Comparing the numerical result of Shooting-VIMHP
obtained as yn ðtÞ ¼ ni¼0 vi ðtÞ. By comparing yVIMHP3 ðtÞjt¼2
with the exact solution of Eq. (23).
with yexact(t)Œt=2 = 1, we found out that the first choice of
A is not appropriate and we should repeat the above procedure t yShootingVIMHP3 yexact jyShootingVIMHP3  yexact j
with second choice of A. By this value of A and new initial 1.0 0.666666 0.666666 0.000000
approximation and finally comparing like powers of p compo- 1.2 0.714294 0.714285 0.000009
nents of solution can be obtained. By comparing the value of 1.4 0.769248 0.769230 0.000018
yVIMHP3 ðtÞjt¼2 that is obtained by this choice of y0 with 1.6 0.833359 0.833333 0.000026
yexact(t)Œt=2 = 1, we found this value of A is not good too. 1.8 0.909121 0.909090 0.000031
2.0 1.000000 1.000000 0.000000
Using these values of A, and applying secant method yields
the best value of A as illustrated in computational algorithm.
The solving procedure by VIMHP with the obtained values
of A in each step is continued until we achieve the acceptable
tolerance. The best value of A that gives the best approxima- Table 2 Comparing the value of yShootingVIMHP3 with the exact
tion of the exact solution in this example is solution at t = 2. for Eq. (23).
A = A4 = .2222644. By studying Fig. 2, it is obvious that n yShootingVIMHP3 yexact jyShootingVIMHP3  yexact j
the obtained solution with the value of A4 has an excellent 1 1.548651000 1 0.548651000
2
agreement with the exact solution yexact ðtÞ ¼ t4 . Graphical 2 1.006730805 1 0.006730805
results that are provided in Fig. 2 show clearly that in compar- 3 1.000341207 1 0.000341207
ison with VIM and VIMHP, Shooting-VIMHP can give the 4 1.000002047 1 0.000002047
best approximation. So, by combining the VIMHP and shoot-
ing method we solved the boundary value problem without
any difficulties and also with less amount of computational 1
work. Table 1 shows the numerical result of the solution y00  y3 þ yy0 ¼ 0; yð1Þ ¼ ; y0 ð1Þ ¼ A; ð34Þ
2
obtained by proposed technique and compares data with the
exact one. Table 2 shows the difference between the result of where the best value of A can be obtained by shooting method
Shooting-VIMHP and the exact solution, where n is the num- and in each step that A is determined in, the problem can be
ber of applying shooting method. By studying Table 2, it is solved by VIMHP. As illustrated in previous example the first
clear that in 4th step for finding the best value of A, we can choice of A is A ¼  16. Using this value and solving Eq. (34) gives
obtain the acceptable solution and our modification causes unacceptable solution because the value of ŒyVIMHP(2) 
the calculations be very simple, the rate of convergence be high yexact(2)Πis large and A should be corrected. Based on computa-
and the number of iteration in VIMHP be decreased. tional algorithm and secant method for finding the best value of
A, after 4th step we can get appropriate value of A and by using
Example 2. Consider the following nonlinear boundary value this value, yVIMHP3 is close enough to the exact solution in each
1
problem, with the exact solutions yðtÞ ¼ tþ1 : point of interval. Using A = .2499 that obtained after 4th step,
our original equation will be converted into:
1 1 1
y00  y3 þ yy0 ¼ 0; yð1Þ ¼ ; yð2Þ ¼ : ð33Þ y00  y3 þ yy0 ¼ 0; yð1Þ ¼ ; y0 ð1Þ ¼ :2499: ð35Þ
2 3 2

For this example, f(t, y, y0 ) = y3  yy0 . So, based on Theo- Applying VIM for finding the best value of Lagrange multi-
rem 2.1, Eq. (33) has unique solution too. To solve this equa- plier gives k(s) = s  t. Using this value and considering the
tion, first of all, we should convert original equation into VIP: VIMHP we have: 8 !3
X1 Z t < X1
n n
p vn ðtÞ ¼ y0 þ p ðs  tÞ  p vn
n¼0 1 : n¼0
! !)
X1 X
1
þ pn vn pn v0n ds; ð36Þ
n¼0 n¼0

and
X
1
yðtÞ ¼ vn ðtÞ:
n¼0

By comparison of like powers of p in order to obtain the

unknown yn, n = 1, 2, 3, . . . we have:

1
p0 : v0 ðtÞ ¼ y0 ¼  0:2499ðt  1Þ
2
Z t
  ðt  1Þ2
1
p : v1 ðtÞ ¼ ðs  tÞ v30 þ v0 v00 ds ¼  ð7:80t3  101t2 þ 596t  1752Þ;
1 10000
Z t
  ðt  1Þ3
p2 : v2 ðtÞ ¼ ðs  tÞ 3v20 v1 þ v0 v01 þ v1 v00 ds ¼ 
1 10000
 ð:020t6  :487t5 þ 5:8t4  430t3 þ 189t2  479t þ 536Þ;
Figure 2 Comparing the result of Shooting-VIMHP, VIMHP ..
. ð37Þ
and VIM with the exact solution of Eq. (23).
Solving BVPs with shooting method and VIMHP 359

1-By this combination there is not any problem to choose

the best initial guess that is necessary in the VIMHP.
2-By the proposed modification there is no need to do huge
and time consuming computational work that is available
in shooting method.
3-By combining these two methods the rate of convergence
is higher.
4-Finally, the excellent agreement between the obtained
results and exact solution approve that applied method is
reliable.

The graphic results and calculations are obtained with the

help of Matlab.

Acknowledgements

Figure 3 Comparing the result of Shooting-VIMHP, VIM and The financial support received from university of Mazandaran
VIMHP with the exact solution of Eq. (33). is gratefully acknowledged. The authors also thank the anon-
ymous reviewer for his/her comments which led to an im-
proved version of the paper.
Table 3 Comparing the numerical result of Shooting-VIMHP
with the exact solution of Eq. (33). References
t yShootingVIMHP3 yexact jyShootingVIMHP3  yexact j [1] L.L. Jiang, T. Zhou, M. Perc, B.H. Wang, Effects of
1.0 0.500000 0.500000 0.000000 competition pattern formation in the rock-paper-scissers game,
1.2 0.454564 0.454545 0.000019 Phys. Rev. E. 84 (2) (2011) (doi:10/1103).
1.4 0.416703 0.416666 0.000037 [2] J. Biazar, P. Gholamian, K. Hosseini, Variational iteration
1.6 0.384666 0.384615 0.000051 method for solving Fokker–Planck equation, J. Franklin Inst.
1.8 0.357190 0.357142 0.000048 347 (7) (2010) 1137–1147.
2.0 0.333333 0.333333 0.000000 [3] Haka K. Akmaz, Variational iteration for elastodynamic
Green’s function, Non. Anal.: Theory Methods Appl. 71 (12)
(2009) 218–223.
[4] J.H. He, Variational iteration method-some recent results and
new interpretations, J. Comput. Appl. Math. 207 (1) (2007) 3–
Table 4 Comparing the value of yShootingVIMHP3 with the exact 17.
solution at t = 2. for Eq. (33). [5] J.S. Nadjafi, A. Ghorbani, He’s homotopy perturbation
n yShootingVIMHP3 yexact jyShootingVIMHP3  yexact j method: an effective tool for solving integral and integro-
differential equation, Comput. Math. Appl 58 (11–12) (2009)
1 0.25759725 0.33333333 0.07573608 2379–2390.
2 0.33369275 0.33333333 0.00035941 [6] J.H. He, Application of homotopy perturbation method to non-
3 0.33333516 0.33333333 0.00000182 linear wave equations, Caos Soliton Fractals 26 (3) (2005) 695–
4 0.33333333 0.33333333 0.00000000 700.
[7] J.H. He, A coupling method of a homotopy technique and a
perturbation technique for non-linear problems, Int. J.
Nonlinear Mech. 35 (1) (2005) 37–43.
By studying Fig. 3, it is clear that there is an excellent agree- [8] G.C. Wu, Adomian decomposition method for non-smooth
ment between yShootingVIMHP3 and the exact solution in compar- initial value problem, Math. Comput. Modell. 54 (9–10) (2011)
ison with VIM and VIMHP. Table 3 shows the numerical 2104–2108.
result for this example and Table 4 as the same as previous [9] A. Hasseine, A. Bellagoun, H.J. Bart, Analytical solution of the
example approves that by applying shooting method and droplet breakup equation by Adomian decomposition method,
introduced algorithm to find the best value of A, the VIMHP App. Math. Comput. 218 (5) (2011) 2249–2258.
can give the excellent approximation. Because, the initial guess [10] N.H. Sweilam, M.M. Khader, Approximate solutions to the
will be better than y0 in the absence of shooting method. nonlinear vibrations of multiwalled carbon nanotubes using
Adomian decomposition method, Appl. Math. Comput. 217 (2)
(2010) 495–505.
4. Conclusion [11] J.H. He, X.H. Wu, Variational iteration method: new
development and applications, Comput. Math. Appl 54 (7-8)
(2007) 881–894.
In this paper we proposed a new technique that is a combina- [12] J. Biazar, H. Ghazvini, M. Eslami, He’s homotopy perturbation
tion of the VIMHP and shooting method to solve boundary method for systems of integro-differential equations, Chaos,
value problems. By this modification we do not encounter Solitons Fractals 39 (2009) 1253–1258.
the difficulties that exist in using the VIMHP and shooting [13] J.H. He, Homotopy perturbation method for solving boundary
method for solving BVPs: value problems, Phys. Lett A 350 (12) (2006) 87–88.
 360
[14] J.H. He, Some asymptotic methods for strongly nonlinear
equations, Int. J. Modern Phys. B 20 (10) (2006) 1141–
1199.
[15] J.H. He, New interpretation of homotopy-perturbation method,
Int. J. Modern Phys. B 20 (2006) 2561–2568.
[16] A. Ghorbani, Beyond Adomian polynomials: He’s polynomials,
Chaos Solitons Fractals 39 (2009) 1486–1492.
[17] S.T. Mahyud-Din, A. Yildirim, Variational iteration method for
delay differential equations using He’s polynomials: Zeitschrif
Fur Naturforschung section A, J. Phys. Sci. 5 (12) (2010) 1045–
1048.
[18] Y. Khan, QB. Wu, Homotopy Perturbation transform method
for nonlinear equations using He’s polynomials, Comput. Math.
Appl. 41 (8) (2011) 1963–1967.
View publication stats
M. Matinfar, M. Ghasemi
[19] S.T. Mahyud-Din, A. Yildirim, Variational iteration method
for Hirota–Satsuma modelusing He’s polynomials: Zeitschrif
Fur Naturforschung section A, J. Phys. Sci. 65 (6–7) (2010)
525–528.
[20] M. Matinfar, M. Ghasemi, Variational iteration method with
He’s polynomials and Laplace transform, J. Nonlinear Sci. Lett.
A 2 (1) (2012) 17–22.
[21] M. Matinfar, M. Ghasemi, Application of variational iteration
method to nonlinear heat transfer equations using He’s
Polynomials. Int. J. Num. Methods Heat Fluid Flow, 23(3)
(2013) 520–531.
[22] H.B. Keller, Numerical Methods for Two-Point Boundary-
Value Problems, Blaisdell, London, 1968, pp.184.