adomian decomposition and modi ed variational iteration ... · pdf fileiteration methods for...
TRANSCRIPT
Adomian Decomposition and Modified Variational
Iteration Methods for Linear Fractional
Integro-Differential Equations
M. H.Saleh, D.Sh.Mohamed,M.H.Ahmed and M.K. Marjan*
Mathematics Department, Faculty of Science,Zagazig University, Zagazig, Egypt
*- E-mail: [email protected]
Abstract
In this paper, we present a comparative study between the modified variational iteration methodand Adomian decomposition method. The study outlines the significant features of the two methods,for solving linear fractional integro-differential equations. From the computational viewpoint, thevariational iteration method is more efficient, convenient and easy to use.
keywords: Fractional integro-differential equations, Adomian decomposition method (ADM),Modified variational iteration method (MVIM), Caputo fractional derivative.
1. Introduction
In recent years various analytical and numerical methods have been applied for the approximate so-lutions of fractional integro-differential equations. Fractional models have been shown by many scientiststo adequately describe the operation of variety of physical and biological processes and systems ([4-7]).Consequently, considerable attention has been given to the solution of fractional integral equations ofphysical interest. Since most fractional integro-differential equations do not have exact analytic solu-tions, approximation and numerical techniques, therefore, are used extensively. Recently, the Adomiandecomposition method ([2],[12]) and Variational iteration method ([10],[11],[13],[14]) have been used forsolving a wide range of problems.To construct an approximate solution to linear integro-differential equations of fractional order. Thevariational iteration method was first proposed by He ([10],[11]) and has been worked out over a numberof years by many authors. This method has been shown to effectively, easily and accurately solve a largeclass of linear problems. Generally, one or two iterations lead to high accurate solutions.In 1980, Adomian proposed decomposition method popularly known as Adomian decomposition method(ADM) has been intensively applied by researchers like Shawagheh (1999,2002), Wazawaz (2001,2005),Odiabat and Momani (2006) as it provides analytical approximate solutions for nonlinear equations with-out linearization, perturbation or discretization. Also Daftardar Gejji and Jafri (2005,2006), Momani andAl-Khaled (2005) provided solution of system of fractional differential equations using Adomian decom-position method [3].Recently Adomian decomposition method is extended for multiterm diffusion-wave equations of fractionalorder(Gejji and (2008))also Mittal et al(2008) solved fractional integro-differential equations by Adomiandecomposition method. The method works very well and solutions of nonlinear problems can be obtainedvery easily and accurately. The method is well reviewed in (Adomian (1988,1992,1994)). The methodinvolving splitting an equation into linear and nonlinear parts and assume that it has a infinite seriessolution. This series has to be truncated for practical purpose but by adding more terms it is possible to
M K Marjan et al, Int.J.Computer Technology & Applications,Vol 6 (2),177-187
IJCTA | Mar-Apr 2015 Available [email protected]
177
ISSN:2229-6093
get arbitrary close to the exact solution in a specific domain.In this paper, we are concerned with the numerical solution of following linear fractional integro-differentialequations:
DαΨ(x) = g(x) +
∫ x
0
k(x, t)Ψ(t)dt, 0 ≤ x, t ≤ 1, (1.1)
with the following supplementary conditions:
Ψ(i)(0) = δi, n− 1 < α ≤ n, n ∈ N , (1.2)
where DαΨ(x) indicates α caputo fractional derivative of Ψ(x), g(x) and k(x, t) are given funtions,x and t are real variables varying in the interval [0, 1], and Ψ(x) is the unknown function to bedetermined.
2. Basic definitions
In this section some basic definitions and properties of fractional calculus theory which are necessaryfor the formulation of the problem are given.
Definition 2.1: A real function f(x), x > 0, is said to be in the space Cµ, µ ∈ R, if there existsa real number p > µ such that f(x) = xp f1(x), where f1(x) ∈ C[0, 1).
Definition 2.2: A function f(x), x > 0, is said to be in the space Cmµ , m ∈ N ∪ {0}, if f (m) ∈ Cµ.
Definition 2.3: The left sided Riemann-Liouville fractional integral operator of order α ≥ 0 of afunction f ∈ Cµ, µ ≥ −1, is defined as [15].
Jαf(x) =1
Γ(α)
∫ x
0
f(t)
(x− t)1−α dt, α > 0, x > 0, (2.1)
J0f(x) = f(x).
Definition 2.4: Let f ∈ Cm−11, m ∈ N ∪ {0}. Then the Caputo fractional derivative of f(x) is de-fined as [4-7]
Dαf(x) =
Jm−αfm(x), m− 1 < α ≤ m, m ∈ N,
Dmf(x)Dxm , α = m.
(2.2)
Hence, we have the following properties:
1. Jα Jνf = Jα+νf, α, ν > 0 , f ∈ Cµ , µ > 0. (2.3)
2. Jα xγ =Γ(γ + 1)
Γ(α+ γ + 1)xα+γ , α > 0, γ > −1, x > 0. (2.4)
3. Jα Dαf(x) = f(x)−m−1∑k=0
f (k)(0+)xk
k!, x > 0, m− 1 < α ≤ m. (2.5)
4. Dα Jα f(x) = f(x), x > 0, m− 1 < α ≤ m. (2.6)
5. Dα C = 0, C is a constant. (2.7)
M K Marjan et al, Int.J.Computer Technology & Applications,Vol 6 (2),177-187
IJCTA | Mar-Apr 2015 Available [email protected]
178
ISSN:2229-6093
6. Dαxβ =
0, β ∈ N0, β < [α],
Γ(β+1)Γ(β−α+1) x
β−α, β ∈ N0, β ≥ [α].
(2.8)
where [α] denoted the smallest integer greater than or equal to α and N0 = {0, 1, 2, . . .}.
3.Analysis of The Modified Variational Iteration Method
In this section, for the convenience of the reader, we first present a brief review of He’s variationaliteration method [16]. Then we will propose the reliable modification of the VIM [1] for solving fractionalintegro-differential equations. Here, we consider the following fractional functional equation
Lu+Ru+Nu = g(x), (3.1)
where L is the fractional order derivative, R is a linear differential operator, N represents the nonlinearterms, and g is the source term. By using (2.5) and applying the inverse operator L−1
x to both sides of(3.1), and using the given conditions, we obtain
u = f − L−1x [Ru]− L−1[Nu] (3.2)
where the function f represents the terms arising from integrating the source term g and from using thegiven conditions, all are assumed to be prescribed.The basic character of He’s method is the construction of a correction functional for (3.1), which reads
un+1(x) = un(x) +
∫ x
0
λ(s)[Lun(s) +Run(s) +Nun(s)− g(s)]ds, (3.3)
where λ is a Lagrange multiplier which can be identified optimally via variational theory [9], un is the nthapproximate solution, and un denotes a restricted variation, i.e., δun = 0.
To solve (3.1) by He’s VIM, we first determine the Lagrange multiplier λ that will be identifiedoptimally via integration by parts. Then the successive approximations un(x), n ≥ 0, of the solutionu(x) will be readily obtained upon using the obtained Lagrange multiplier and by using any selectivefunction u0. The approximation u0 may be selected by any function that just satisfies at least the initialand boundary conditions. With determined λ, then several approximations un(x), n ≥ 0, follow im-mediately. Consequently, the exact solution may be obtained by using
limn−→1
un(x) = u(x). (3.4)
In summary, we have the following variational iteration formula for (3.2){u0(x) is an arbitrary initial guess,
un+1(x) = un(x) +∫ x
0λ(s)[Lun(s) +Run(s) +Nun(s)− g(s)]ds,
(3.5)
or equivalently, for (3.2), according to [16]:{u0(x) is an arbitrary initial guess,
un+1(x) = f(x)− L−1x [Ru]− L−1
x [Nu],(3.6)
where the multiplier Lagrange λ, has been identified.It is important to note that He’s VIM suggests that the u0 usually defined by a suitable trial-function
with some unknown parameters or any other function that satisfies at least the initial and boundaryconditions. This assumption made by He ([8],[9]) and others will be slightly varied, as will be seen in thediscussion.
The MVIM, that was introduced by Ghorbani et al [1], can be established based on the assumption
M K Marjan et al, Int.J.Computer Technology & Applications,Vol 6 (2),177-187
IJCTA | Mar-Apr 2015 Available [email protected]
179
ISSN:2229-6093
that the function f(x) of the iterative relation (3.6) can be divided into two parts, namely f0(x) andf1(x). Under this assumption, we set
f(x) = f0(x) + f1(x). (3.7)
According to the assumption, (3.7), and by the relationship (3.6), we construct the following variationaliteration formula
u0(x) = f0(x),
u1(x) = f(x)− L−1x [Rf0(x)]− L−1
x [Nf0(x)],
un+1(x) = f(x)− L−1x [Run(x)]− L−1
x [Nun(x)],
where the multiplier Lagrange λ, has been identified. Here, a proper selection was proposed for thecomponents u0(x) and u1(x). The suggestion was that only the part f0 be assigned to the zerothcomponent i.e u0. An important observation that can be made here is that the success of the proposedmethod depends mainly on the proper choice of the functions f0 and f1. As will be seen from theexamples below, this selection of u0 will result in a reduction of the computational work and acceleratethe convergence. Furthermore, this proper selection of the components u0 and u1 may provide the solutionby using one iteration only. To give a clear overview of the content of this study. We have chosen severalfractional integro-differential equations with the nonlocal boundary conditions.
4.Analysis of The Adomian Decomposition Method
Consider the equation (1.1) with initial condition (1.2) where Dα is the operator defined as (2.2).Operating with Jα on both sides of the equation (1.1) as follows:
Ψ(x) =
m−1∑k=0
Ψ(k)(0+)xk
k!+ Jαg(x) + Jα
∫ x
0
k(x, t)Ψ(t)dt, 0 ≤ x, t ≤ 1, (4.1)
Adomian decomposition method defines the solution by the series:
Ψ(x) =∞∑i=0
Ψi(x), (4.2)
where the components Ψ0(x),Ψ1(x),Ψ2(x), ...,Ψi(x), ... are determined recursively by
Ψ0(x) =
m−1∑k=0
Ψ(k)(0+)xk
k!+ Jαg(x),
Ψi+1(x) = Jαg(x) + Jα∫ x
0
k(x, t)Ψi(t)dt, (4.3)
decomposition method suggests that 0th component Ψ0(x) be defined by the initial conditions and thefunction g(x) as described above. The other components namely Ψ0(x),Ψ1(x), . . .etc. are derived recur-rently.
5. Numerical examples
In this section, some numerical examples of linear fractional integro-differential equations are pre-sented to illustrate the above results. All results are obtained by using Maple 16.
Example 5.1Consider the following fractional integro-differential equation:
D57 Ψ(x) =
2401
276
x237
Γ( 27 )− 1
5x5 sin(x)− x sin(x) +
∫ x
0
sin(x)Ψ(t)dt, (5.1)
M K Marjan et al, Int.J.Computer Technology & Applications,Vol 6 (2),177-187
IJCTA | Mar-Apr 2015 Available [email protected]
180
ISSN:2229-6093
subject to Ψ(0) = 1 with the exact solution Ψ(x) = 1 + x4.
By assuming L = D57 and applying the inverse operator L−1
x to both sides of (5.1), we have
Ψ(x) = 1 + J57 [
2401
276
x237
Γ( 27 )
]− J 57 [
1
5x5 sin(x) + x sin(x)] + J
57
∫ x
0
sin(x)Ψ(t)dt. (5.2)
According to the original VIM (3.5), and corresponding the recursive scheme (3.6) and by using (3.7),we obtain
g(x) = g0(x) + g1(x) = 1 + J57 [ 2401
276x
237
Γ( 27 )
]− J 57 [ 1
5x5 sin(x) + x sin(x)],
= 1 + x4 − J 57 [ 1
5x5 sin(x) + x sin(x)],
by assuming
g0(x) = 1 + x4, g1(x) = −J 57 [
1
5x5 sin(x) + x sin(x)], (5.3)
and with starting of the initial approximation Ψ0(x) = g0(x) = 1 + x4, we have
Ψ0(x) = 1 + x4,
Ψ1(x) = 1 + x4 − J 57 [ 1
5x5 sin(x) + x sin(x)] + L−1
x [Ψ0(x)],
Ψ1(x) = 1 + x4 − J 57 [
1
5x5 sin(x) + x sin(x)] + J
57 [
1
5x5 sin(x) + x sin(x)] = 1 + x4, (5.4)
Ψn+1(x) = 1 + x4 − J 57 [ 1
5x5 sin(x) + x sin(x)] + L−1
x [Ψn(x)]⇒ Ψn+1(x) = 1 + x4, n ≥ 0.
In view of (5.4), we obtain the approximate solution Ψ(x) = 1 + x4, which is the same of the exactsolution.
Now applying Adomian decomposition method for equation (5.2) by using equation (4.2) we get:
∞∑i=0
Ψi(x) = 1 + J57 [
2401
276
x237
Γ( 27 )
]− J 57 [
1
5x5 sin(x) + x sin(x)] + J
57
∫ x
0
sin(x)∞∑i=0
Ψi(t)dt, (5.5)
by using equation (5.5) we obtain as the follows:
Ψ0(x) = 1 + J57 g(x)
Ψ0(x) = 1 + x4 − 0.07160086639x477 + 0.007058904187x
617 − 0.0003033879182x
757 + . . . ,
Ψ1(x) = J57λ∫ x
0k(x, t)Ψ0(t)dt,
Ψ1(x) = 0.05264769589x477 − 0.007058904186x
617 + 0.0003033879182x
757 + . . . ,
Ψ2(x) = J57λ∫ x
0k(x, t)Ψ1(t)dt,
Ψ2(x) = −5.773754204× 10−12x1697 + 1.089352272× 10−13x
1837 − 1.315207672× 10−15x
1977 + . . . ,
Ψ3(x) = J57λ∫ x
0k(x, t)Ψ2(t)dt,
Ψ3(x) = −2.129464558× 10−10x1607 + 1.152235978× 10−11x
1747 − 3.718717854× 10−24x
2727 + . . . ,
hance∑∞i=0 Ψi(x) = Ψ0 + Ψ1 + Ψ2 + Ψ3 + . . . .
The approximation solution for α = 5/7 is
Ψ(x) = 1+x4−2.129464558×10−10x1607 +1.152235978×10−11x
1747 −3.718717854×10−24x
2727 − . . . .
M K Marjan et al, Int.J.Computer Technology & Applications,Vol 6 (2),177-187
IJCTA | Mar-Apr 2015 Available [email protected]
181
ISSN:2229-6093
Table 1 and figure 1 show the comparison between Modified variational method and Adomian Decompo-sition method, we find that Modified variational method is better than Adomian decomposition method.
Table 1x Exact Approx variational Approx adomian Errer variational Errer adomian0 1 1 1 0 0
0.1 1.0001 1.0001 1.000099996 0 3.659466400× 10−9
0.2 1.0016 1.0016 1.001599616 0 3.843611821× 10−7
0.3 1.0081 1.0081 1.008094147 0 0.0000058526144810.4 1.0256 1.0256 1.025559568 0 0.000040431724990.5 1.0625 1.0625 1.062318800 0 0.00018120002540.6 1.1296 1.1296 1.128982201 0 0.00061779906990.7 1.2401 1.2401 1.238355374 0 0.0017446263650.8 1.4096 1.4096 1.405307042 0 0.0042929571950.9 1.6561 1.6561 1.646589249 0 0.0095107504601.0 2.0000 2.0000 1.980601618 0 0.01939838302
Figure 1: Numerical results of Example 5.1.
M K Marjan et al, Int.J.Computer Technology & Applications,Vol 6 (2),177-187
IJCTA | Mar-Apr 2015 Available [email protected]
182
ISSN:2229-6093
Example 5.2Consider the following fractional integro-differential equation:
D23 Ψ(x) =
3
4
x13
√3 Γ( 2
3 )(3x− 2)
π− 1
4x4 exp(x)− 1
3x3 exp(x) +
∫ x
0
exp(x) tΨ(t)dt, (5.6)
subject to Ψ(0) = 0 with the exact solution Ψ(x) = x2 − x.By assuming L = D
23 and applying the inverse operator L−1
x to both sides of (5.6), we have
Ψ(x) = 0 + J23 [
3
4
x13
√3 Γ( 2
3 )(3x− 2)
π]− J 2
3 [1
4x4 exp(x) +
1
3x3 exp(x)] + J
23
∫ x
0
exp(x) t Ψ(t)dt, (5.7)
According to the original VIM (3.5), and corresponding the recursive scheme (3.6) and by using (3.7),we obtain
g(x) = g0(x) + g1(x) = 0 + J23 [ 3
4
x13√
3 Γ( 23 )(3x−2)
π ]− J 23 [ 1
4x4 exp(x) + 1
3x3 exp(x)],
= x2 − x− J 23 [ 1
4x4 exp(x) + 1
3x3 exp(x)],
by assuming
g0(x) = x2 − x, g1(x) = −J 23 [
1
4x4 exp(x) +
1
3x3 exp(x)], (5.8)
and with starting of the initial approximation Ψ0(x) = g0(x) = x2 − x, we have
Ψ0(x) = x2 − x,
Ψ1(x) = x2−x−J 23 [
1
4x4 exp(x)+
1
3x3 exp(x)]+L−1
x [Ψ0(x)], (5.9)
Ψ1(x) = x2 − x− J 34 [ 1
4x4 exp(x) + 1
3x3 exp(x)] + J
34 [ 1
4x4 exp(x) + 1
3x3 exp(x)] = x2 − x,
Ψn+1(x) = x2 − x− J 34 [ 1
4x4 exp(x) + 1
3x3 exp(x)] + L−1
x [Ψn(x)]⇒ Ψn+1(x) = x2 − x, n ≥ 0.
In view of (5.9), we obtain the approximate solution Ψ(x) = x2 − x, which is the same of the exactsolution.
Now applying Adomian decomposition method for equation (5.7) by using equation (4.2), we get:
∞∑i=0
Ψi(x) = 0 + J23 [
3
4
x13
√3 Γ( 2
3 )(3x− 2)
π]− J 2
3 [1
4x4 exp(x) +
1
3x3 exp(x)] + J
23
∫ x
0
exp(x)t∞∑i=0
Ψi(t)dt,
(5.10)by using equation (5.10) we obtain as the follows:
Ψ0(x) = J23 g(x),
Ψ0(x) = x2 − x+ 0.02913191738x14/3 − 0.02570463301x17/3 − 0.01927847474x20/3 − . . . ,
Ψ1(x) = J23λ∫ x
0k(x, t)Ψ0(t)dt,
Ψ1(x) = 0.02570463300x17/3 + 0.01927847474x20/3 + 0.01056125138x23/3 − 0.02913191740x14/3 + . . . ,
Ψ2(x) = J23λ∫ x
0k(x, t)Ψ1(t)dt,
Ψ2(x) = −0.006885695970x19/3−0.007399887556x22/3−0.003123602081x25/3+0.0005423582779x34/3+ . . . ,
Ψ3(x) = J23λ∫ x
0k(x, t)Ψ2(t)dt,
Ψ3(x) = 4.873687408×10−8x19+0.000007245575142x16+0.00001413636375x15−0.0003449563982x10+ . . . ,
M K Marjan et al, Int.J.Computer Technology & Applications,Vol 6 (2),177-187
IJCTA | Mar-Apr 2015 Available [email protected]
183
ISSN:2229-6093
hance∑∞i=0 Ψi(x) = Ψ0 + Ψ1 + Ψ2 + Ψ3 + . . . .
The approximation solution for α = 2/3 is
Ψ(x) = x2−x−2×10−11x14/3−1×10−11x17/3−6×10−13x28/3−1×10−13x37/3−5×10−13x31/3 +. . . .
Table 2 and figure 2 show the comparison between Modified variational method and Adomian Decompo-sition method, we find that Modified variational method is better than Adomian decomposition method.
Table 2x Exact Approx variational Approx adomian Errer variational Errer adomian0 0 0 0 0 0
0.1 -0.09 -0.09 -0.09 0 9.999999987× 10−25
0.2 -0.16 -0.16 -0.16 0 2.047983077× 10−21
0.3 -0.21 -0.21 -0.2100000000 0 1.768962184× 10−19
0.4 -0.24 -0.24 -0.2399999998 0 2.000000040× 10−10
0.5 -0.25 −0.25 -0.2499999965 0 3.600000003× 10−9
0.6 -0.24 -0.24 -0.2399999619 0 3.799999695× 10−8
0.7 -0.21 -0.21 -0.2099997080 0 2.918998698× 10−7
0.8 -0.16 -0.16 -0.1599982462 0 0.0000017536968700.9 -0.09 -0.09 -0.08999127985 0 0.0000087201486691.0 0 0 0.00003727337387 0 0.00003727337387
Figure 2: Numerical results of Example 5.2.
M K Marjan et al, Int.J.Computer Technology & Applications,Vol 6 (2),177-187
IJCTA | Mar-Apr 2015 Available [email protected]
184
ISSN:2229-6093
Example 5.3Consider the following fractional integro-differential equation:
D35 Ψ(x) =
125
14
x125 sin( 2
5π) Γ( 35 )
π− 1
3x
132 +
∫ x
0
√x t2Ψ(t)dt, (5.11)
subject to Ψ(0) = 0 with the exact solution Ψ(x) = 2x3.
By assuming L = D35 and applying the inverse operator L−1
x to both sides of (5.11), we have
Ψ(x) = 0 + J35 [
125
14
x125 sin( 2
5π) Γ( 35 )
π]− J 3
5 [1
3x
132 ] + J
35
∫ x
0
√x t2Ψ(t)dt, (5.12)
According to the original VIM (3.5), and corresponding the recursive scheme (3.6) and by using (3.7),we obtain
g(x) = g0(x) + g1(x) = 0 + J35 [ 125
14
x125 sin( 2
5π) Γ( 35 )
π ]− J 35 [ 1
3x132 ]
= 2x3 − J 35 [ 1
3x132 ]
by assuming
g0(x) = 2x3, g1(x) = −J 35 [
1
3x
132 ], (5.13)
and with starting of the initial approximation Ψ0(x) = g0(x) = 2x3, we have
Ψ0(x) = 2x3,
Ψ1(x) = 2x3 − J 35 [
1
3x
132 ] + L−1
x [Ψ0(x)], (5.14)
Ψ1(x) = 2x3 − J 35 [ 1
3x132 ] + J
35 [ 1
3x132 ] = 2x3,
Ψn+1(x) = 2x3 − J 35 [ 1
3x132 ] + L−1
x [Ψn(x)]⇒ Ψn+1(x) = 2x3, n ≥ 0.
In view of (5.14), we obtain the approximate solution Ψ(x) = 2x3, which is the same of the exact solution.
Now applying Adomian decomposition method for equation (5.12) by using equation (4.2), we get:
∞∑i=0
Ψi(x) = 0 + J35 [
125
14
x125 sin( 2
5π) Γ( 35 )
π]− J 3
51
3x
132 + J
35
∫ x
0
√x t2
∞∑i=0
Ψi(t)dt, (5.15)
by using equation (5.15) we obtain as the follows:
Ψ0(x) = J35 g(x),
Ψ0(x) = 2x3 − 0.1011008935x71/10,
Ψ1(x) = J35λ∫ x
0k(x, t)Ψ0(t)dt,
Ψ1(x) = −0.002324006305x56/5 + 0.1011008935x71/10,
Ψ2(x) = J35λ∫ x
0k(x, t)Ψ1(t)dt,
Ψ2(x) = −0.00003160247042x153/10 + 0.002324006305x56/5,
Ψ3(x) = J35λ∫ x
0k(x, t)Ψ2(t)dt,
Ψ3(x) = −2.896664654× 10−7x97/5 + 0.00003160247042x153/10,
hance∑∞i=0 Ψi(x) = Ψ0 + Ψ1 + Ψ2 + Ψ3 + . . . .
M K Marjan et al, Int.J.Computer Technology & Applications,Vol 6 (2),177-187
IJCTA | Mar-Apr 2015 Available [email protected]
185
ISSN:2229-6093
The approximation solution for α = 3/5 is
Ψ(x) = 2x3 − 2.896664654× 10−7x97/5.
Table 3 and figure 3 show the comparison between Modified variational method and Adomian Decompo-sition method, we find that Modified variational method is better than Adomian decomposition method.
Table 3x Exact Approx variational Approx adomian Errer variational Errer adomian0 0 0 0 0 0
0.1 0.002 0.002 002 0 1.153182970× 10−26
0.2 0.016 0.016 0.016 0 7.977744732× 10−21
0.3 0.054 0.054 0.054 0 2.079938824× 10−17
0.4 0.128 0.128 0.128 0 5.519021066× 10−15
0.5 0.250 0.250 0.250 0 4.187128641× 10−13
0.6 0.432 0.432 0.4320000000 0 1.438906178× 10−11
0.7 0.686 0.686 0.6859999997 0 2.862860564× 10−10
0.8 1.024 1.024 1.023999996 0 3.818070713× 10−9
0.9 1.458 1.458 1.457999962 0 3.751482313× 10−8
1.0 2.000 2.000 1.999999710 0 2.896664654× 10−7
Figure 3: Numerical results of Example 5.3.
M K Marjan et al, Int.J.Computer Technology & Applications,Vol 6 (2),177-187
IJCTA | Mar-Apr 2015 Available [email protected]
186
ISSN:2229-6093
ConclusionIn this paper we study the numerical solution of three examples by using Modified variational iteration
method and Adomian decomposition method which derive a good approximation. The study shows thatthe techniques require less computational work than existing approaches while supplying quantitativelyreliable results.
References
[1] A. Ghorbani, J. Saberi-Nadjafi, An effective modification of He’s variational iterationmethod, Nonlinear Analysis: Real World Applications 10 (2009) 2828-2833.
[2] A.M. Wazwaz, The modified Adomian decomposition method for solving linear and nonlinearboundary value problems of tenth-order and twelfth-order, Int. J. Nonlinear Sci. Numer.Simul. 1 (2000) 17-24.
[3] Babolian E., Davari A., Numerical implementation of Adomian decomposition method; Ap-plied Mathematics and Computing. 153 (2004), 301-305.
[4] I. Podlubny, Fractional Differential Equations, Academic Press, New York, 1999. method,Chaos, Solitons and Fractals (2007).
[5] I. Podlubny, Geometric and physical interpretation of fractional integration and fractionaldifferentiation, Fract. Calc. Appl. Anal. 5 (2002) 367-386.
[6] J.H. He, Nonlinear oscillation with fractional derivative and its applications. An: Interna-tional Conference on Vibrating Engineering 98, Dalian, China, 1998, pp. 288-291.
[7] J. H. He, Some applications of nonlinear fractional differential equation and their approxi-mations, Bull. Sci. Technol. 15 (12) (1999) 86-90.
[8] J. H. He, Variational iteration method for delay differential equations, Commun. NonlinearSci. Numer. Simul. 2 (4) (1997) 235-236.
[9] J. H. He, Variational iteration method-Some recent results and new interpretations, J. Com-put. Appl. Math, 207 (2007) 3-17.
[10] J.H. He, Variational iteration methoda kind of non-linear analytical technique: some exam-ples. Int. J. Non-Linear Mech. 34 (1999) 699-708.
[11] J.H. He, Variational iteration method for autonomous ordinary differential systems. Appl.Math. Comput. 114 (2000) 115-123.
[12] M. Rahman, Integral Equations and their Applications, Britain Athenaeum Press Ltd.(2007).
[13] S. Momani, Z. Odibat, Numerical comparison of methods for solving linear differential equa-tions of fractional order, Chaos Solitons Fractals 31 (2007) 1248-1255.
[14] S. Irandoust-Pakchin and S. Abdi-Mazraeh, Exact solutions for some of the fractionalintegro-differential equations with the nonlocal boundary conditions by using the modifica-tion of He’s variational iteration method, International Journal of Advanced MathematicalSciences 1(3)(2013), 139-144.
[15] S. Miller, B. Ross; An Introduction to the Fractional Calculus and Fractional DifferentialEquations, John Wiley and Sons, USA, 1993.
[16] Yongxiang Zhao, Aiguo Xiao; Variational iteration method for solving perturbation initialvalue problems; Computer physics communications. Vol(181), iss(5), (2012), 947-956.
M K Marjan et al, Int.J.Computer Technology & Applications,Vol 6 (2),177-187
IJCTA | Mar-Apr 2015 Available [email protected]
187
ISSN:2229-6093