Variational iteration method and homotopy perturbation method are used to solve the fractional Fredholm integrodifferential equations with constant coefficients. The obtained results indicate that the method is efficient and also accurate.
1. Introduction
The topic of fractional calculus has attracted many scientists because of its several applications in many areas, such as physics, chemistry, and engineering. For a detail survey with collections of applications in various fields, see, for example, [1–3].
Further, the fractional derivatives technique has been employed for solving linear fractional differential equations including the fractional integrodifferential equations; in this way, much of the efforts is devoted to searching for methods that generate accurate results, see [4, 5]. In this work, we present two different methods, namely, homotopy perturbation method and variational iteration method [6], for solving a fractional Fredholm integro-differential equations with constant coefficients. There is a vast literature, and we only mention the works of Liao which treat a homotopy method in [7, 8].
For the nonlinear equations with derivatives of integer order, many methods are used to derive approximation solution [9–14]. However, for the fractional differential equation, there are some limited approaches, such as Laplace transform method [3], the Fourier transform method [15], the iteration method [16], and the operational calculus method [17].
Recently, there has been considerable researches in fractional differential equations due to their numerous applications in the area of physics and engineering [18], such as phenomena in electromagnetic theory, acoustics, electrochemistry, and material science [3, 16, 18, 19]. Similarly, there is also growing interest in the integrodifferential equations which are combination of differential and Fredholm-Volterra equations. In this work, we study these kind of equations that have the fractional order usually difficult to solve analytically, thus a numerical method is required, for example, the successive approximations, Adomian decomposition, Chebyshev and Taylor collocation, Haar Wavelet, Tau and Walsh series methods.
This note is devoted to the application of variational iteration method (VIM) and homotopy perturbation method (HPM) for solving fractional Fredholm integrodifferential equations with constant coefficients:
∑k=0∞PkD*αy(t)=g(x)+λ∫0aH(x,t)y(t)dt,a≤x,t≤b,
under the initial-boundary conditions
D*αy(a)=y(0),D*αy(0)=y′(a),
where a is constant and 1<α<2 and D*α is the fractional derivative operator given in the Caputo sense. For the physical understanding of the fractional integrodifferential equations, see [20]. Further, we also note that fractional integrodifferential equations were associated with a certain class of phase angles and suggested a new way for understanding of Riemann’s conjecture, see [21].
Outline of this paper is as follows. Section 2 contains preliminaries on fractional calculus. Section 3 is a short review of the homotopy method and Section 4 variational iteration method. Sections 5 and 6 are devoted to VIM and HPM analysis, respectively. Concluding remarks with suggestions for future work are listed in Section 7.
2. Description of the Fractional Calculus
In the following, we give the necessary notations and basic definitions and properties of fractional calculus theory; for more details, see [3, 13, 16, 22].
Definition 2.1.
A real function f(x),x>0, is said to be in the space Cα, α∈R if there exists a real number (p>α), such that f(x)=xpf1(x), where f1(x)=C([0,∞)). Clearly, Cα⊂Cβ, if β≤α.
Definition 2.2.
A function f(x),x>0, is said to be in space Cαm, m∈N, if f(m)∈Cα.
Definition 2.3.
The Riemann-Liouville fractional integral of order μ≥0for a function f∈Cα,(α≥1) is defined as
Iμf(t)=1Γ(μ)∫0t(t-τ)μ-1f(τ)dτ,μ>0,t>0,
in particular I0f(t)=f(t).
Definition 2.4.
The Caputo fractional derivative of f∈C-1m, m∈N, is defined as
Dcμf(t)={[Im-μf(m)(t)],m-1<μ≤m,m∈N,dmdtmf(t),μ=m.
Note that
The Riemann-Liouville fractional integral operator of order ρ≥0 for a function f∈Cμ,(μ≥-1) is defined as
Kρf(x)=1Γ(ρ)∫0x(x-t)ρ-1f(t)dt,ρ>0,x>0,K0f(x)=f(x),
having the properties
KρKβf(x)=Kρ+βf(x),Kρxβ=Γ(β+1)Γ(ρ+β+1)xα+β.
According to the Caputo’s derivatives, we obtain the following expressions:
CDμC=0,C=constant,CDμtβ={Γ(β+1)Γ(α+β+1)tβ-α,β>α-1,0,β≤α-1.
Lemma 2.6.
If m-1<α≤m, m∈N, f∈Cμm, μ≥-1, then the following two properties hold:
(1)DαKαf(t)=f(t),(2)(DαKα)f(t)=f(t)-∑k=0m-1f(k)(0+)tkk!.
In fact, Kılıçman and Zhour introduced the Kronecker convolution product and expanded to the Riemann-Liouville fractional integrals of matrices by using the Block Pulse operational matrix as follows:
1Γ(α)∫0t(t-t1)α-1ϕm(t1)dt1≃Fαϕm(t),
whereFα=(bm)α1Γ(α+2)[1ξ2ξ3⋯ξm01ξ2⋯ξm-1001⋯ξm-2000⋱⋮00001],
see [23].
In our work, we consider Caputo fractional derivatives and apply the homotopy method in order to derive an approximate solutions of the fractional integrodifferential equations.
3. Homotopy Method
To illustrate the basic ideas of this method, we consider the following nonlinear differential equation:
A(u)+f(r)=0,r∈Ω,
with boundary conditions
B(u,∂u∂n)=0,r∈Γ,
where A is a general differential operator, B is a boundary operator, f(r) is a known analytical function, and Γ is the boundary of the domain Ω, see [24].
In general, the operator A can be divided into two parts L and N, where L is linear, while N is nonlinear. Equation (3.1), therefore, can be rewritten as follows:L(u)+N(u)-f(r)=0.
By using the homotopy technique that was proposed by Liao in [7, 8], we construct a homotopy of (3.1) v(r,p):Ω×[0,1]→ℛ which satisfiesH(v,p)=(1-p)[L(v)-L(u0)]+p[A(v)+f(r)]=0,p∈[0,1],r∈Ω,
or
H(v,p)=L(v)-L(u0)+pL(u0)+p[N(v)-f(r)]=0,
where p∈[0,1] is an embedding parameter and u0 is an initial approximation which satisfies the boundary conditions. By using (3.4) and (3.5), we have
H(v,0)=L(v)-L(u0)=0,H(v,1)=A(v)-f(r)=0.
The changing in the process of p from zero to unity is just that of v(r,p) from u0 to u(r). In a topology, this is also
known deformation, further L(v)-L(u0) and A(v)-f(r) are homotopic.
Now, assume that the solution of (3.4) and (3.5) can be expressed asv=v+pv1+p2v2+⋯.
The approximate solution of (3.1), therefore, can be readily obtained:u=limp→1v=v0+v1+v2+⋯.
The convergence of the series of (3.8) has been proved in the [25, 26].
4. The Variational Iteration Method
To illustrate the basic concepts of the VIM, we consider the following differential equation:Lu+Nu=g(x),
where L is a linear operator, N is a nonlinear operator, and g(x) is an nonhomogenous term; for more details, see [19].
According to the VIM, one construct a correction functional as follows:un+1(x)=un(x)+∫0xλ[Lun(s)+Nũn(s)-g(s)]ds,
where λ is a general Lagrange multiplier, which can be identified optimally via the variational theory, and the subscript n denotes the order of approximation, ũu is considered variation [6, 27], that is, δũu=0.
5. Analysis of VIM
To solve the fractional integrodifferential equation (1.1) by using the variational iteration method, with boundary conditions (1.2), one can construct the following correction functional:
yk+1(x)=yk(x)+∫0tμ∑k=0∞PkD*αy(s)ds-μg̃k(x)-λ∫abμH(x,s)ỹk(s)ds,
where μ is a general Lagrange multiplier and g̃k(x) and ỹk(x) are considered as restricted variations, that is, δg̃k(x)=0 and δỹk(x)=0.
Making the above correction functional stationary, the following conditions can be obtained:
δyk+1(x)=δyk(x)+∫0t[∑k=0∞Pkμ(s)δD*αy(s)-δg̃k(x)-λ∫abH(x,s)μ(s)δỹk(s)ds],
having the boundary conditions as follows:1-μ′(s)∣x=s=0,μ(s)∣x=s=1.
The Lagrange multipliers can be identified as follows:μ(s)=12(s-x).
Substituting the value of μ from (5.4) into correction functional of (5.1) leads to the following iteration formulae:yk+1(x)=yk(x)+μ2Γ(α-1)∫0x(x-s)α-2(s-x)×[∫0t∑k=0∞PkD*αy(s)ds-g̃k(x)-λ∫abH(x,s)ỹk(s)]ds,yk+1(x)=yk(x)-μ(α-1)2Γ(α)∫0x(x-s)α-1×[∫0t∑k=0∞PkD*αy(s)ds-g̃k(x)-λ∫abH(x,s)ỹk(s)]ds,
by applying formulae (2.4), we getyk+1(x)=yk(x)-(α-1)Kα2Γ(α)[∫0tμ∑k=0∞PkD*αy(v)dv-μg̃k(x)-λ∫abμH(x,v)ỹk(v)]dv.
The initial approximation can be chosen in the following manner which satisfies initial boundary conditions (1.2)-(1.3):y0(x)=υ0+υ1x,whereυ1=D*αy(0),υ0=D*αy(a).
We can obtain the following first-order approximation by substitution of (5.7) into (5.6)y1(x)=y0(x)-(α-1)Kα2Γ(α)[∫0tμ∑k=0NPkD*αy(v)dv-μg̃0(x)-λ∫abμH(x,v)ỹk(v)]dv.
Substituting the constant value of υ0 and υ1 in the expression (5.8) results in the approximation solution of (1.1)–(1.3).
6. Analysis of HPM
This section illustrates the basic of HPM for fractional Fredholm integrodifferential equations with constant coefficients (1.1) with initial-boundary conditions (1.2).
In view of HPM [25, 26], construct the following homotopy for (1.1):∑k=0∞PkD*αy(x)=p[∑k=0∞PkD*αy(x)+(g(t)-λ∫abH(x,t)y(x)dx)].
In view of basic assumption of HPM, solution of (1.1) can be expressed as a power series in p:y(x)=D*αy0(x)+pD*αy1(x)+p2D*αy2(x)+p3D*αy3(x)+⋯.
If we put p→1 in (6.2), we get the approximate solution of (1.1):y(x)=D*αy0(x)+D*αy1(x)+D*αy2(x)+D*αy3(x)+⋯.
The convergence of series (6.3) has been proved in [28].
Now, we substitute (6.2) into (6.1); then equating the terms with identical power of p, we obtain the following series of linear equations:p0:∑k=0∞PkD*αy0(t)=0,p1:∑k=0∞PkD*αy1(t)=∑k=0∞PkD*αy0(t)-λ∫abH(x,t)y0(x)dx,p2:∑k=0∞PkD*αy2(t)=∑k=0∞PkD*αy1(t)+g(x)-λ∫abH(x,t)y1(x)dx,p3:∑k=0∞PkD*αy3(t)=∑k=0∞PkD*αy2(t)-λ∫abH(x,t)y2(x)dx,p4:∑k=0∞PkD*αy4(t)=∑k=0∞PkD*αy3(t)-λ∫abH(x,t)y3(x)dx,
with the initial-boundary conditionsD*αy(a)=y(0),D*αy(0)=y′(a).
We can also take the initial approximation in the following manner which satisfies initial-boundary conditions (1.2)-(1.3):y0(x)=υ0+υ1x,whereυ1=D*αy(0),υ0=D*αy(a).
Note that (6.4) can be solved by applying the operator Kβ, which is the inverse of operator Dα we approximate the series solution of HPM by the following n-term truncated series [29]:χn(x)=D*αy0(x)+D*αy1(x)+D*αy2(x)+D*αy3(x)+⋯+D*αyn-1(x),
which results, the approximate solutions of (1.2)-(1.3). For further analysis, the variational iteration method, see [30] and the algorithm by the homotopy perturbation method, see [31].
7. Conclusion
The proposed methods are used to solve fractional Fredholm integrodifferential equations with constant coefficients. Comparison of the results obtained by the present method with that obtained by other method reveals that the present method is very effective and convenient. Unfortunately, the disadvantage of the second method is that the embedding parameter p is quite casual, and often enough the approximations obtained by this method will not be uniform. So, in our future work we expect to study this kind of equation by using a combination of the variational iteration method and the homotopy perturbation method which has shown reliable results in supplying analytical approximation that converges very rapidly. However, we note that the papers [32, 33] suggest alternative ways for similar problems.
Acknowledgments
The first author would like to thank Professor Juan J. Trujillo for the very helpful discussion. The authors also acknowledge that this research was partially supported by University Putra Malaysia under the Research University Grant Scheme 05-01-09-0720RU.
HilfertR.2000SingaporeWorld Scientificviii+46310.1142/97898128177471890104OldhamK. B.SpanierJ.1974198Academic Pressxiii+234Mathematics in Science and Engineering0361633PodlubnyI.1999198San Diego, Calif, USAAcademic Pressxxiv+340Mathematics in Science and Engineering1658022KademA.BaleanuD.Fractional radiative transfer equation within Chebyshev spectral approach20105951865187310.1016/j.camwa.2009.08.0302595960ZBL1189.35359KademA.BaleanuD.Analytical method based on Walsh function combined with orthogonal polynomial for fractional transport equation201015349150110.1016/j.cnsns.2009.05.0242572186ZBL1221.45008HeJ. H.Variational iteration method for delay differential equations1997242352362-s2.0-000699094710.1016/S1007-5704(97)90008-3LiaoS. J.An approximate solution technique not depending on small parameters: a special example199530337138010.1016/0020-7462(94)00054-E1336915ZBL0837.76073LiaoS. J.Boundary element method for general nonlinear differential operators199720291992-s2.0-0031223353AblowitzM. J.ClarksonP. A.1991149New York, NY, USACambridge University Pressxii+51610.1017/CBO97805116239981149378ZBL0900.65350AdomianG.198946Dordrecht, The NetherlandsKluwer Academic Publishersxx+224982493ZBL0754.60066ChenY.YanZ.Weierstrass semi-rational expansion method and new doubly periodic solutions of the generalized Hirota-Satsuma coupled KdV system20061771859110.1016/j.amc.2005.10.0372234499ZBL1094.65104FanE.Extended tanh-function method and its applications to nonlinear equations20002774-521221810.1016/S0375-9601(00)00725-81827770ZBL1167.35331LuchkoY.GorenfloR.An operational method for solving fractional differential equations with the Caputo derivatives19992422072331710779ZBL0931.44003MomaniS.OdibatZ.Numerical approach to differential equations of fractional order200720719611010.1016/j.cam.2006.07.0152332951ZBL1119.65127KempfleS.BeyerH.Global and causal solutions of fractional differential equationsProceedings of the 2nd International Workshop on Transform Methods and Special Functions1996Varna, BulgariaScience Culture Technology Publishing210216SamkoS. G.KilbasA. A.MarichevO. I.1993Yverdon, SwitzerlandGordon and Breachxxxvi+9761347689LuchkoYu. F.SrivastavaH. M.The exact solution of certain differential equations of fractional order by using operational calculus1995298738510.1016/0898-1221(95)00031-S1327564ZBL0824.44011WestB. J.BolognaM.GrigoliniP.2003New York, NY, USASpringerx+3541988873HeJ. H.Variational iteration method—a kind of non-linear analytical technique: some examples19993446997082-s2.0-0000092673He.J. H.ElaganS. K.LiZ. B.Geometrical explanation of the fractional complex transform and derivative chain rule for fractional calculus20123764257259Le MéhautéA.El KaabouchiA.NivanenL.Contribution of non integer Integro-differential operators(NIDO) to the geometrical undersanding of Riemann’s conjecture-(I)2Proceedings of the 2nd IFAC workshop on Fractional Differentiation and Its Applications2006230233KilbasA. A.SrivastavaH. M.TrujilloJ. J.2006204Amsterdam, The NetherlandsElsevier Sciencexvi+523North-Holland Mathematics Studies2218073KılıçmanA.Al ZhourZ. A. A.Kronecker operational matrices for fractional calculus and some applications2007187125026510.1016/j.amc.2006.08.1222323577ZBL1123.65063HeJ. H.Homotopy perturbation method: a new nonlinear analytical technique20031351737910.1016/S0096-3003(01)00312-51934316ZBL1030.34013HeJ. H.Homotopy perturbation technique19991783-425726210.1016/S0045-7825(99)00018-31711041ZBL0956.70017HeJ. H.A coupling method of a homotopy technique and a perturbation technique for non-linear problems2000351374310.1016/S0020-7462(98)00085-71723761HeJ. H.Some asymptotic methods for strongly nonlinear equations200620101141119910.1142/S02179792060337962220565ZBL1102.34039HeJ. H.2006Berlin, GermanyGmbHNawazY.Variational iteration method and homotopy perturbation method for fourth-order fractional integro-differential equations2011618233023412-s2.0-7995374850410.1016/j.camwa.2010.10.004HeJ. H.A short remark on fractional variational iteration method2011375383362336410.1016/j.physleta.2011.07.0332826245Ji-HuanH. E.A Note on the homotopy perturbation method20101425655682-s2.0-77955858024YildirimA.SezerS. A.KaplanY.Numerical solutions of fourth-order fractional integro-differential equations20106512102710322-s2.0-78649835796YüzbasiS.SahinN.YildirimA.A collocation approach for solving high-order linear Fredholm-Volterra integro-differential equations2012553-4547563