IJERA (International journal of Engineering Research and Applications) is International online, ... peer reviewed journal. For more detail or submit your article, please visit www.ijera.com
1. H. Saeedi, F. Samimi / International Journal of Engineering Research and Applications
(IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 2, Issue 5, September- October 2012, pp.052-056
Heโs Homotopy Perturbation Method for Nonlinear Ferdholm
Integro-Differential Equations Of Fractional Order
H. Saeedi *, F. Samimi **
*(Department of Mathematics, Shahid Bahonar University of Kerman, Iran
** (Department of Mathematics, Islamic Azad University of Zahedan, Iran
ABSTRACT
In this paper, homotopy perturbation method is devoted to the search for better and more efficient
applied to solve non-linear Fredholm integro- solution methods for determining a solution,
differential equations of fractional order. approximate or exact, analytical or numerical, to
Examples are presented to illustrate the ability of nonlinear models. So the aim of this work is to
the method. The results reveal that the proposed present a numerical method (Homotopy-perturbation
methods very effective and simple and show that method) for approximating the solution of a
this method can be applied to the non-fractional nonlinear fractional integro-differential equation of
cases. the second kind:
1
๐ท๐ผ ๐ ๐ฅ โ ๐ ๐ ๐ฅ, ๐ก ๐น ๐ ๐ก ๐๐ก
Keywords - Fractional calculus, Fredholm 0
integro-differential equations, Homotopy- = ๐(๐ฅ) (1)
perturbation method Where ๐น(๐ ๐ก ) = ๐ ๐ก ๐ ๐>1
With these supplementary conditions:
I. INTRODUCTION ๐ (๐) 0 = ๐ฟ ๐ , ๐ = 0,1, โฆ , ๐ โ 1,
The generalization of the concept of derivative ๐โ1 < ๐ผ โค ๐ , ๐
๐ท ๐ผ ๐(๐ฅ) to non-integer values of a goes to the โ ๐ (2)
beginning of the theory of differential calculus. In Where,๐ โ ๐ฟ2 ([0, 1)), ๐ โ ๐ฟ2 ([0,1)2 ) are known
fact, Leibniz, in his correspondence with Bernoulli, functions, f(x) is the unknown function, ๐ท ๐ผ is the
LโHopital and Wallis (1695), had several notes about Caputo fractional differentiation operator and q is a
1
the calculation of ๐ท 2 ๐(๐ฅ). Nevertheless, the positive integer. There are several techniques for
development of the theory of fractional derivatives solving such equations like Adomian decomposition
and integrals is due to Euler, Liouville and Abel method [7, 9], collocation method [8], CAS wavelet
(1823). However, during the last 10 years fractional method [2] and differential transform method [1].
calculus starts to attract much more attention of Most of the methods have been utilized in linear
physicists and mathematicians. It was found that problems and a few numbers of works have
various; especially interdisciplinary applications can considered nonlinear problems. In this paper,
be elegantly modeled with the help of the fractional homotopy perturbation method is applied to solve
derivatives. For example, the nonlinear oscillation of non-linear Fredholm integro-differential equations of
earthquake can be modeled with fractional fractional order.
derivatives [3], and the fluid-dynamic traffic model
with fractional derivatives caneliminate the II. FRACTIONAL CALCULUS
deficiency arising from the assumption of continuum There are several definitions of a fractional derivative
traffic flow [4]. In the fields of physics and of order ๐ผ> 0. The two most commonly used
chemistry, fractional derivatives and integrals are definitions are the RiemannโLiouville and Caputo.
presently associated with the application of fractals in Definition1.2.The Riemann-Liouville fractional
the modeling of electro-chemical reactions, integral operator of order ฮฑ โฅ 0 is defined as
irreversibility and electro magnetism [10], heat
conduction in materials with memory and radiation ๐ฝ๐ผ ๐ ๐ฅ
๐ฅ
problems. Many mathematical formulations of 1
mentioned phenomena contain nonlinear integro- = (๐ฅ โ ๐ก) ๐ผ โ1 ๐ ๐ก ๐๐ก, ๐ผ > 0, ๐ฅ
ฮ(๐ผ) 0
differential equations with fractional order. Nonlinear > 0, 3
phenomena are also of fundamental importance in ๐ฝ0 ๐ ๐ฅ = ๐(๐ฅ)
various fields of science and engineering. The
nonlinear models of real-life problems are still IIt has the following properties:
difficult to be solved either numerically or
theoretically. There has recently been much attention
52 | P a g e
2. H. Saeedi, F. Samimi / International Journal of Engineering Research and Applications
(IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 2, Issue 5, September- October 2012, pp.052-056
ฮ ๐พ+1 The solution, usually an approximation to the
๐ฝ๐ผ ๐ฅ๐พ = ๐ฅ ๐ผ+๐พ , solution, will be obtained by taking the limit as p
ฮ ๐ผ+ ๐พ+1
๐พ > โ1, (4) tends to 1,
Definition2.2.The Caputo definition of fractal ๐ = lim ๐ = ๐0 + ๐1 + ๐2
๐โ1
derivative operator is given by +โฏ (9)
To illustrate the homotopy perturbation method, for
๐ท ๐ผ ๐ ๐ฅ = ๐ฝ ๐โ๐ผ ๐ท ๐ ๐ ๐ฅ nonlinear Ferdholm integro-differential equations of
๐ฅ
1 fractional order, we consider
= ๐ฅ
ฮ(๐ โ ๐ผ) 0
โ ๐ก ๐ โ๐ผโ1 ๐ ๐ (๐ก)๐๐ก, (5) 1โ ๐ ๐ท ๐ผ ๐๐ ๐ฅ
1
Where, m โ 1 โค ฮฑ โค m, mโN, x >0. It has the + ๐ ๐ท ๐ผ ๐๐ ๐ฅ โ ๐ ๐ ๐ฅ โ ๐ ๐ ๐ฅ, ๐ก [๐๐ ๐ก ] ๐ ๐๐ก
0
following two basic properties:
=0 10
Or
๐ท๐ผ ๐ฝ๐ผ ๐ ๐ฅ = ๐ ๐ฅ ,
๐ฝ๐ผ ๐ท๐ผ ๐ ๐ฅ ๐ท ๐ผ ๐๐ ๐ฅ = ๐ ๐๐ ๐ฅ
= ๐ ๐ฅ 1
๐ โ1
๐ +
๐ฅ๐ + ๐ ๐ ๐ฅ, ๐ก [๐๐ ๐ก ] ๐ ๐๐ก (11)
โ ๐ (0 ) , 0
๐!
๐ =0 P is an embedding parameter which changes from
๐ฅ>0 (6) zero to unity. Using the parameter p, we expand the
solution of the Eq. (1) in the following form:
๐๐ ๐ก = ๐0 + ๐๐1 + ๐2 ๐2 + โฏ. (12)
III. HOMOTOPY PERTURBATION METHOD Substituting (12) into (11) and collecting the terms
FOR SOLVING EQ. (1) with the same powers of p, we obtain a series of
To illustrate the basic concept of homotopy equations of the form:
perturbation method, consider the following non- ๐0 : ๐ท ๐ผ ๐0 = 0 (13)
linear functional equation: 1
๐1 : ๐ท๐ผ ๐ = ๐ ๐ฅ + ๐
1 ๐ ๐ฅ, ๐ก ๐0 ๐ก ๐
๐๐ก (14)
0
๐ด ๐ = ๐ ๐ (7) 1
With the following boundary conditions: ๐2 : ๐ท ๐ผ ๐2 = ๐ ๐ ๐ฅ, ๐ก ๐ ๐ก
1
๐
๐๐ก (15)
0
๐๐ .
๐, = 0, ๐โฮ .
๐๐
Where A is a general functional operator, B is a It is obvious that these equations can be easily solved
boundary operator, f(r) is a known analytic function, by applying the operator ๐ฝ ๐ผ , the inverse of the
and ฮ is the boundary of the domain ฮฉ. Generally operator๐ท ๐ผ , which is defined by (3). Hence, the
speaking the operator A can be decomposed into two components ๐๐ (i = 0, 1, 2 , ...) of the HPM solution
parts L and N, where L is a linear and N is a non- can be determined. That is, by setting p = 1. In (11)
linear operator. Eq. (7), therefore, can be rewritten as we can entirely determine the HPM series
the following: solutions, ๐ ๐ฅ = โ ๐๐ ๐ฅ .
๐=0
๐ฟ ๐ + ๐ ๐ โ ๐ ๐ =0 IV. APPLICATIONS
In this section, in order to illustrate the method, we
We construct a homotopy ๐ ๐, ๐ : ฮฉ ร 0,1 โ ๐ , solve two examples and then we will compare the
which satisfies obtained results with the exact solutions.
๐ป ๐, ๐ = 1 โ ๐ ๐ฟ ๐ โ ๐ฟ ๐0 Example 4.1. Consider the following fractional
+ ๐ ๐ด ๐ โ ๐ ๐ = 0, nonlinear integro-differential equation:
1
๐ โ 0,1 , ๐โฮฉ (8) 32 2
๐ท๐ผ ๐ ๐ฅ = 2โ ๐ฅ + ๐ฅ 2 ๐ก 2 ๐ ๐ก 2 ๐๐ก,
Or 15 0
๐ป ๐, ๐ = ๐ฟ ๐ โ ๐ฟ ๐0 + ๐๐ฟ ๐0 0 โค ๐ฅ < 1,
+ ๐ ๐ ๐ โ ๐ ๐ =0 0< ๐ผโค1 (16)
Where, ๐ข0 is an initial approximation for the solution With this supplementary condition ๐(0) = 1.
of Eq. (7).In this method, we use the homotopy For ๐ผ = 1, we can get the exact solution
parameter p to expand ๐ ๐ฅ = 1 + 2๐ฅ.
1
For ๐ผ = , According to (11) we construct the
๐ = ๐0 + ๐๐1 + ๐2 ๐2 + โฏ 4
following homotopy
53 | P a g e
3. H. Saeedi, F. Samimi / International Journal of Engineering Research and Applications
(IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 2, Issue 5, September- October 2012, pp.052-056
1
1
๐ท 4 ๐ ๐ฅ = ๐(๐ ๐ฅ + ๐ฅ2 ๐ก 2 ๐ ๐ก 2
๐๐ก) (17)
0
Substituting (12) into (17)
1
๐0 โถ ๐ท 4 ๐0 ๐ฅ = 0
1
1
๐1 โถ ๐ท 4 ๐ ๐ฅ = ๐ ๐ฅ +
1 ๐ฅ 2 ๐ก 2 [๐ ๐ก ]2 ๐๐ก
0
0
1
1
๐2 : ๐ท 4 ๐ ๐ฅ =
2 ๐ฅ 2 ๐ก 2 [2๐0 ๐ก ๐1 ๐ก ]๐๐ก
0
1
3 1
๐ โถ ๐ท 4 ๐3 ๐ฅ = ๐ฅ 2 ๐ก 2 [2๐0 ๐ก ๐2 ๐ก + ๐12 ๐ก ]๐๐ก
0
1
1
๐4 โถ ๐ท 4 ๐ ๐ฅ =
4 ๐ฅ 2 ๐ก 2 [2๐0 ๐ก ๐3 ๐ก
0
+ 2๐2 ๐ก ๐ ๐ก ]๐๐ก
1
.
.
.
Consequently, by applying the operators ๐ฝ ๐ผ to the
above sets Example 4.2. Consider the following nonlinear
Fredholm integro-differential equation, of order
1
๐0 ๐ฅ = 1 ๐ผ=2
1
1 2
1154 1 764 9 ๐ท 2 ๐ ๐ฅ = ๐ ๐ฅ + ๐ฅ๐ก ๐ ๐ก ๐๐ก
๐ =
1 ๐ฅ 4โ ๐ฅ 4 0
523 541 0โค ๐ฅ<1 (18)
893 9
๐2 (๐ฅ) = 1386 ๐ฅ 4 1 17
Where, ๐ ๐ฅ = 3 ๐ฅ โ 12 ๐ฅ with these
718 9 ฮ( )
2
๐3 ๐ฅ = ๐ฅ 4 supplementary conditions๐ 0 = 1, with the exact
1169
317 9 solution, ๐ ๐ฅ = 1 + ๐ฅ
๐4 ๐ฅ = 1250 ๐ฅ 4
According to (11) we construct the following
. homotopy:
. 1
1 2
. ๐ท 2 ๐ ๐ฅ = ๐(๐ ๐ฅ + ๐ฅ๐ก ๐ ๐ก ๐๐ก) (19)
Therefore the approximations to the solutions of 0
1 Substituting (12) into (18)
Example 4.1. For ๐ผ = 4 will be determined as 1
๐0 โถ ๐ท 2 ๐0 ๐ฅ = 0
1
โ 1
๐ ๐ฅ = ๐๐ ๐ฅ = ๐0 + ๐ + ๐2 + ๐3 + ๐4 + โฏ
1
๐1 โถ ๐ท 2 ๐ ๐ฅ = ๐ ๐ฅ +
1 ๐ฅ๐ก[๐0 ๐ก ]2 ๐๐ก
๐=0 0
1
1154 4 1
๐ฅ = 1+ ๐2 โถ ๐ท 2 ๐2 ๐ฅ = ๐ฅ๐ก[2๐0 ๐ก ๐ ๐ก ]๐๐ก
1
523 0
1 1
The approximations to the solutions for ๐ผ = is 1
2821 2
2 ๐3 โถ ๐ท 2 ๐3 ๐ฅ = ๐ฅ๐ก[2๐0 ๐ก ๐2 ๐ก + ๐12 ๐ก ]๐๐ก
๐ ๐ฅ = 1+ ๐ฅ. 0
1250 1
3 1
Hence for ๐ผ = 4 , ๐4 โถ ๐ท 2 ๐4 ๐ฅ = ๐ฅ๐ก[2๐0 ๐ก ๐3 ๐ก
346 4 0
๐ ๐ฅ = 1+ ๐ฅ3. + 2๐2 ๐ก ๐ ๐ก ]๐๐ก
159 1
1 1 3
Fig. 1 shows the numerical results for ๐ผ = . . . .
4 2 4
.
The comparisons show that as ๏ก ๏ฎ 1 , the .
approximate solutions tend to ๐(๐ฅ) = 1 + 2๐ฅ, Consequently, by applying the operators ๐ฝ ๐ผ to the
which is the exact solution of the equation in the case above sets
of ๐ผ = 1. ๐0 ๐ฅ = 1
1379 3
๐ = ๐ฅโ
1 ๐ฅ 2
2000
41 3
๐2 (๐ฅ) = ๐ฅ 2
200
54 | P a g e
4. H. Saeedi, F. Samimi / International Journal of Engineering Research and Applications
(IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 2, Issue 5, September- October 2012, pp.052-056
617 3 232 7
๐3 ๐ฅ = ๐ฅ 2 ๐3 ๐ฅ = ๐ฅ 4
5269 921
471 3 779 7
๐4 ๐ฅ = ๐ฅ 2 ๐4 ๐ฅ = 5000 ๐ฅ 4
6173
. .
. .
. .
Therefore the approximations to the solutions of
Example 4.2. Will be determined as Therefore the approximations to the solutions of
โ
Example 4.3. Will be determined as
๐ ๐ฅ = ๐๐ ๐ฅ = ๐0 + ๐ + ๐2 + ๐3 + ๐4 + โฏ
1 โ
๐=0 ๐ ๐ฅ = ๐๐ ๐ฅ = ๐0 + ๐1 + ๐2 + ๐3 + ๐4 + โฏ
=1+ ๐ฅ ๐=0
And hence,๐ ๐ฅ = 1 + ๐ฅ, which are the exact = 1 + ๐ฅ2
solution. And hence ๐ ๐ฅ = 1 + ๐ฅ 2 , which are the exact
Example 4.3. Consider the following nonlinear solution.
Fredholm integro-differential equation, of order
3
๐ผ=4 V. CONCLUSION
This paper presents the use of the Heโs homotopy
3 8 5 15 perturbation method, for non-linear ferdholm
๐ท 4 ๐ ๐ฅ = 5
๐ฅ 4โ ๐ฅ integro-differential equations of fractional order. We
5ฮ 8
4 usually derive very good approximations to the
1
3
solutions. It can be concluded that the Heโs homotopy
+ ๐ฅ๐ก ๐ ๐ก ๐๐ก perturbation method is a powerful and efficient
0
0โค ๐ฅ<1 (20) technique infinding very good solutions for this kind
of equations.
With these supplementary Condition ๐ 0 = 1, with
the exact solution, ๐ ๐ฅ = 1 + ๐ฅ 2
According to (11) we construct the following ACKNOWLEDGEMENTS
homotopy: The authors wish to thank the reviewers for their
3 useful Comments on this paper.
๐ท 4๐ ๐ฅ
1
3
= ๐(๐ ๐ฅ + ๐ฅ๐ก ๐ ๐ก ๐๐ก) (21)
0
Substituting (12) into (21) REFERENCES
3 [1] Erturk VS, Momani S, Odibat Z.
๐0 โถ ๐ท 4 ๐0 ๐ฅ = 0 Application of generalized differential
1
3 transform method to multi-order fractional
๐1 โถ ๐ท 4 ๐ ๐ฅ = ๐ ๐ฅ +
1 ๐ฅ๐ก[๐0 ๐ก ]3 ๐๐ก
0 differential equations. Commun Nonlinear
1
3 Sci Numer Simulat,(2008), 13:1642โ54.
๐2 โถ ๐ท 4 ๐2 ๐ฅ = ๐ฅ๐ก[3๐0 ๐ก 2
๐ ๐ก ]๐๐ก
1 [2] H. Saeedi, M. M. Moghadam, N.
0
3
1 Mollahasani, and G. N. Chuev, A CAS
๐3 โถ ๐ท 4 ๐3 ๐ฅ = ๐ฅ๐ก[3๐0 ๐ก 2
๐2 ๐ก wavelet method for solving nonlinear
0 Fredholm Integro-differential equations of
+ 3๐0 ๐ก ๐ ๐ก 2 ]๐๐ก
1 fractional order, Communications in
1
3 Nonliner Science and Numerical Simulation,
๐4 โถ ๐ท 4 ๐4 ๐ฅ = ๐ฅ๐ก[3๐0 ๐ก 2
๐3 ๐ก + 6๐0 ๐ก
0 vol. 16, no. 3, (2011) pp. 1154โ1163..
+ ๐1 ๐ก + ๐2 ๐ก + ๐32 (๐ก)]๐๐ก [3] He JH. Nonlinear oscillation with fractional
. derivative and its applications . In:
. International conference on vibrating
. engineeringโ98. China: Dalian: (1998), p.
Consequently, by applying the operators ๐ฝ ๐ผ to the 288โ91.
above sets [4] He JH. "Some applications of nonlinear
๐0 ๐ฅ = 1 fractional differential equations and their
approximations". Bull Sci Techno;15(2),
2137 7 (1999), 86โ90.
๐ = ๐ฅ2 โ
1 ๐ฅ 4
[5] J.H. He, Compute.Methods Appl. Mech.
2500
Eng. (1999), 178.257.
203 7
๐2 (๐ฅ) = 479 ๐ฅ 4
55 | P a g e
5. H. Saeedi, F. Samimi / International Journal of Engineering Research and Applications
(IJERA) ISSN: 2248-9622 www.ijera.com
Vol. 2, Issue 5, September- October 2012, pp.052-056
[6] J.H. He, Int. J. Nonlinear Mech. (2000),
35.37.
[7] Momani S, Noor M. Numerical methods for
fourth order fractional integro-differential
equations. Appl Math Comput, 182: (2006),
754โ60.
[8] Rawashdeh E. Numerical solution of
fractional integro-differential equations by
collocation method. Appl Math
Comput,176:16, 2006 .
[9] Ray SS. Analyticalsolution for the space
fractional diffusion equation by two-step
adomian decomposition method. Commun
Nonlinear SciNumerSimulat, 14: (2009),
1295โ306.
[10] Tenreiro Machado JA. Analysis and design
of fractional-order digital control systems.
Sys Anal Modell Simul;107โ22.1997.
56 | P a g e