We introduce two new spectral wavelets algorithms for solving linear and nonlinear fractional-order Riccati differential equation. The suggested algorithms are basically based on employing the ultraspherical wavelets together with the tau and collocation spectral methods. The main idea for obtaining spectral numerical solutions depends on converting the differential equation with its initial condition into a system of linear or nonlinear algebraic equations in the unknown expansion coefficients. For the sake of illustrating the efficiency and the applicability of our algorithms, some numerical examples including comparisons with some algorithms in the literature are presented.
1. Introduction
The fractional calculus deals with derivatives and integrals to an arbitrary order (real or complex order). Fractional ordinary differential equations are crucial in many fields. For example, many physical phenomena in acoustics, damping laws and electroanalytical chemistry, neuron modeling, diffusion processing, and material sciences (see, e.g., [1–3]) are described by differential equations of fractional order.
The use of fractional differential and integral operators in mathematical models has gained great importance in recent years. Several forms of fractional differential equations have been proposed in standard models, and there has been significant interest in developing numerical schemes for their solutions. There are numerous methods that deal with these types of equations; some of these methods are Laplace transform [4], Adomian decomposition method (ADM) [5], variational iteration method (VIM) [6], fractional difference method (FDM) [7], a quadrature tau method [8], and a shifted Jacobi spectral method [9].
Spectral methods are a class of techniques used extensively in various fields such as applied mathematics and scientific computing (see, e.g., [10–12]). These methods aim to find numerical solutions to different kinds of differential equations. They are used in solving ordinary, partial, and fractional differential equations. When spectral methods are used, the numerical solution is assumed as an expansion in terms of certain special functions which are called “basis functions.” A great advantage of spectral methods is that they take on a global approach unlike finite element methods which use a local approach. There are three well-known versions of spectral methods, namely, tau, collocation, and Galerkin methods (see, e.g., [13]). The choice of the suitable spectral method suggested for solving the given equation depends certainly on the type of the differential equation and on the type of the boundary conditions governed by it.
The ultraspherical polynomials have received considerable attention in recent decades, from both theoretical and practical points of view (see, e.g., [14]). Some authors are interested in employing these polynomials for solving various kinds of differential equations. In this respect, Elgindy and Smith-Miles in [15] treated boundary value problems and integral and integrodifferential equations using ultraspherical integration matrices. Moreover, Doha and Abd-Elhameed employed ultraspherical polynomials for solving one- and two-dimensional second-order differential equations [16]. In addition, the same authors in [17] developed some accurate spectral solutions for the parabolic and elliptic partial differential equations based on the ultraspherical tau method.
The use of wavelets in solving ordinary and fractional differential equations has attracted a large number of authors. For example, in a huge number of articles, Legendre and Chebyshev wavelets are employed for treating ordinary differential equations as well as fractional differential equations (see, e.g., [18–20]). To the best of our knowledge, no articles deal with the application of ultraspherical wavelets in handling fractional-order differential equations. This motivates our interest in such kind of wavelets hoping to employ them in various practical applications. Another motivation is that the Chebyshev and Legendre wavelets can be deduced as special cases of the ultraspherical wavelets.
The fractional-order Riccati differential equation is one of the important fractional-order differential equations. This equation has applications in various disciplines, such as random processes, optimal control, and diffusion problems. We refer the interested readers in the fundamental theories of this equation and its applications to the book of Reid [21].
The main objective in this paper is to introduce two efficient spectral algorithms based on using ultraspherical wavelets for solving fractional-order Riccati differential equation. The two algorithms, namely, ultraspherical wavelets tau method (UWTM) and ultraspherical wavelets collocation method (UWCM), are implemented and presented for handling both of linear and nonlinear fractional-order Riccati differential equations.
The paper is organized as follows. Section 2 is concerned with presenting some necessary definitions and mathematical preliminaries of the fractional calculus theory which are required for establishing our results. Also, in this section, some properties of ultraspherical polynomials and their shifted ones are presented. Section 3 is devoted to constructing the ultraspherical wavelets. In Section 4, some new spectral wavelets solutions for fractional-order Riccati differential equations are developed based on employing the ultraspherical wavelets which were constructed in Section 3. In Section 5, the convergence of the suggested ultraspherical wavelets expansion is carefully investigated. Some numerical examples and discussions are given in Section 6 aiming to illustrate the efficiency, simplicity, and applicability of the suggested algorithms. Some conclusions are given in Section 7.
2. Preliminaries2.1. Basic Definitions of Fractional Calculus
This section is concerned with presenting some notations, definitions, and preliminary facts of the fractional calculus theory which will be useful throughout this paper.
Definition 1.
The Riemann-Liouville fractional integral operator Iα of order α on the usual Lebesgue space L1[0,1] is defined as
(1)Iαf(t)={1Γ(α)∫0t(t-τ)α-1f(τ)dτ,α>0,f(t),α=0.
The operator Iα has the following properties:
(2)(i)IαIβ=Iα+β,(ii)IαIβ=IβIα,(iii)(IαIβf)(t)=(IβIαf)(t),(iv)Iα(t-a)ν=Γ(ν+1)Γ(ν+α+1)(t-a)ν+α,
where f∈L1[0,1], α,β⩾0, and ν>-1.
Definition 2.
The Riemann-Liouville fractional derivative of order α>0 is defined by
(3)(Dαf)(t)=(ddt)n(In-αf)(t),n-1⩽α<n,n∈N.
The following well-known properties for the operator Dα are useful:
(4)DαC=0,Cisaconstant,Dαtβ={0,forβ∈N0,β<⌈α⌉,Γ(β+1)Γ(β+1-α)tβ-α,forβ∈N0,β≥⌈α⌉,
where the ceiling notation ⌈α⌉ denotes the smallest integer greater than or equal to α and N0={0,1,2,…}. Recall that, for α∈N, the fractional-differential operator coincides with the usual differential operator of integer-order. Similar to integer-order differentiation, also Dα is a linear operator; that is,
(5)Dα(λf(t)+μg(t))=λDαf(t)+μDαg(t),
where λ and μ are constants.
Definition 3.
The Caputo definition of fractional differential operator is given by
(6)(D*αf)(t)=1Γ(n-α)∫0t(t-τ)n-α-1f(n)(τ)dτ,α>0,t>0,
where n-1⩽α<n, n∈N.
The operator D*α satisfies the following two basic properties for n-1⩽α<n:(7)(D*αIαf)(t)=f(t),(IαD*αf)(t)=f(t)-∑k=0n-1f(k)(0+)k!(t-a)k,t>0.
For more details on mathematical properties of fractional derivatives and integrals, see, for example, [4, 22].
2.2. Some Properties of Ultraspherical Polynomials and Their Shifted Ones
The ultraspherical polynomials (a special type of Jacobi polynomials) associated with the real parameter (λ>-1/2) are a sequence of orthogonal polynomials defined on (−1, 1), with respect to the weight function w(x)=(1-x2)λ-1/2. The orthogonality relation is given by
(8)∫-11(1-x2)λ-1/2Cm(λ)(x)Cn(λ)(x)dx={πn!Γ(λ+1/2)(2λ)n(n+λ)Γ(λ),m=n,0,m≠n,
where
(9)(2λ)n=Γ(n+2λ)Γ(2λ).
It should be noted here that the ultraspherical polynomials Cn(λ)(x) are normalized such that Cn(λ)(1)=1. This normalization is characterized by an advantage that the polynomials Cn(0)(x) are identical with the Chebyshev polynomials of the first kind Tn(x), Cn(1/2)(x) are the Legendre polynomials Ln(x), and Cn(1)(x) is equal to (1/(n+1))Un(x), where Un(x) are the Chebyshev polynomials of the second kind.
The polynomials Cn(λ)(x) may be generated by using the recurrence relation
(10)(n+2λ)Cn+1(λ)(x)=2(n+λ)xCn(λ)(x)-nCn-1(λ)(x),n=1,2,3,…,
with the initial values C0(λ)(x)=1 and C1(λ)(x)=x.
For more properties and relations of ultraspherical polynomials, see, for instance, [23].
The shifted ultraspherical polynomials C~n(λ)(x)=Cn(λ)(2x-1) are a sequence of orthogonal polynomials defined on (0,1), with respect to the weight function w~(x)=(x-x2)λ-1/2; that is,
(11)∫01(x-x2)λ-1/2C~m(λ)(x)C~n(λ)(x)dx={π21-4λΓ(n+2λ)n!(n+λ)(Γ(λ))2,m=n,0,m≠n.
They also may be generated by using the recurrence relation
(12)(n+2λ)C~n+1(λ)(x)=2(n+λ)(2x-1)C~n(λ)(x)-nC~n-1(λ)(x),n=1,2,3,…,
with the initial values C~0(λ)(x)=1 and C~1(λ)(x)=2x-1.
All relations and properties of ultraspherical polynomials can be easily transformed to give the corresponding relations and properties of the shifted ultraspherical polynomials.
Now, the following integral formula (see [23]) is needed:
(13)∫Cn(λ)(x)w(x)dx=-2λ(1-x2)λ+1/2n(n+2λ)Cn-1(λ+1)(x),n⩾1.
Also, the following theorem is essential in investigating the convergence analysis for the suggested ultraspherical wavelets expansion.
Theorem 4 (Bernstein-type inequality; see [24]).
The following inequality holds for ultraspherical polynomials:
(14)(sinθ)λ|Cn(λ)(cosθ)|<21-λΓ(n+3λ/2)Γ(λ)Γ(n+1+λ/2),0⩽θ⩽π,0<λ<1.
3. Construction of Ultraspherical Wavelets
Wavelets constitute a family of functions constructed from dilation and translation of single function called the mother wavelet. When the dilation parameter a and the translation parameter b vary continuously, we have the following family of continuous wavelets:
(15)ψa,b(t)=|a|-1/2ψ(t-ba)a,b∈R,a≠0.
We define the ultraspherical wavelets ψnm(λ)(t)=ψ(k,n,m,λ,t) as they have five arguments: k,n can be assumed to be any positive integer, m is the order for the ultraspherical polynomial, λ is the known ultraspherical parameter, and t is the normalized time. Explicitly, they are defined on the interval [0,1] as
(16)ψnm(λ)(t)={2k/2ξmCm(λ)(2k-1t-n+1),t∈[n-12k-1,n2k-1],0,otherwise,
where m=0,1,…,M-1, n=1,2,…,2k-1, and
(17)ξm=2λΓ(λ)m!(m+λ)2πΓ(m+2λ).
Remark 5.
It is worth noting here that ψnm(1/2)(t) is identical to Legendre wavelets [25, 26], ψnm(0)(t) is identical to first kind Chebyshev wavelets [27, 28], and ψnm(1)(t) is identical to second kind Chebyshev wavelets [29, 30].
Now, consider a function f(t) defined on [0,1] and suppose that f(t) may be expanded in terms of the ultraspherical wavelets as
(18)f(t)=∑n=1∞∑m=0∞cnmψnm(λ)(t),
where
(19)cnm=(f(t),ψnm(λ)(t))w~=∫01(t-t2)λ-1/2f(t)ψnm(λ)(t)dt.
Also, we can assume that the function f(t) can be approximated in terms of ultraspherical wavelets as
(20)f(t)≈∑n=12k-1∑m=0M-1cnmψnm(λ)(t).
4. Spectral Solutions of Fractional-Order Riccati Differential Equation
In this section, with the aid of the ultraspherical wavelets expansion together with the two well-known spectral methods, namely, tau and collocation methods, we numerically solve the following fractional-order Riccati differential equation:
(21)Dαu(t)=p(t)u(t)2+q(t)u(t)+r(t),t∈(0,1),0<α⩽1,
subject to the initial condition
(22)u(0)=B.
The following lemma is needed for developing the wavelets solutions of fractional-order Riccati differential equation.
Lemma 6.
The fractional derivative of order α∈(ν-1,ν) of the ultraspherical wavelets (16) is given by
(23)Dαψnm(λ)(t)={21+(α-ν+1/2)kξmΓ(α-ν)(ddt)ν×∫02k-1t-n+1C~m(λ)(τ)(2k-1t-n+1-τ)α-ν+1dτ,t∈[n-12k-1,n2k-1],0,hhotherwise.
Proof.
The result can be immediately obtained by applying the operator defined in (3) on the ultraspherical wavelets ψnm(λ)(t) defined in (16).
4.1. Case I: Linear Fractional-Order Riccati Differential Equation
In this subsection, we consider (21) in case of p(t)=0; that is, our aim is to discuss the following fractional-order Riccati linear differential equation:
(24)Dαu(t)=q(t)u(t)+r(t),t∈(0,1),0<α⩽1,
subject to the initial condition (22).
Now, assume that the approximate solution of (24) can be expanded in terms of the ultraspherical wavelets as
(25)uk,M,λ(t)=∑n=12k-1∑m=0M-1cnmψnm(λ)(t),
and if we substitute (25) into (24), then the residual of this equation is given by
(26)R(t)=∑n=12k-1∑m=0M-1cnm(Dαψnm(λ)(t)-q(t)ψnm(λ)(t))-r(t).
The application of the typical tau method (see, e.g., [13]) implies that
(27)∫01R(t)ψnm(λ)(t)dt=0,(n,m)∈Λ,
where
(28)Λ={1,2,…,2k-1}×{0,1,…,M-1}-{(2k-1,M-1)}.
Moreover, the use of the initial condition (22) yields
(29)∑n=12k-1∑m=0M-1cnmψnm(λ)(0)=B.
Equations (27) and (29) generate a set consisting of (2k-1M) equations in the unknown expansion coefficients, {cnm:1≤n≤2k-1;0≤m≤M-1}. This system of equations can be solved with the aid of any suitable solver such as Gauss-Seidel method (see, [31]). Hence, the unknown components of the vector C can be obtained, and accordingly the required spectral wavelets solution uk,M,λ(t) given by (25) can be obtained.
4.2. Case II: Nonlinear Fractional-Order Riccati Differential Equation
In this subsection, we aim to handle the following nonlinear fractional-order Riccati differential equation:
(30)Dαu(t)=p(t)u(t)2+q(t)u(t)+r(t),t∈(0,1),0<α⩽1,p(t)≠0,
subject to the initial condition (22).
Assume that the approximate solution of (30) can be expanded in terms of the ultraspherical wavelets as in (25). Substitution of (25) into (30) enables one to write the residual of (30) as
(31)R(t)=∑n=02k-1∑m=0M-1cnmDαψnm(λ)(t)-p(t)(∑n=02k-1∑m=0M-1cnmψnm(λ)(t))2hhhhhhhh-q(t)(∑n=02k-1∑m=0M-1cnmψnm(λ)(t))-r(t).
The application of the collocation method (see, e.g., [13]) implies that
(32)R(ti)=0,i=1,2,…,2k-1M-1,
where the points ti, i=1,2,…,2k-1M-1 are selected to be the first (2k-1M-1) roots of C2k-1M(λ)(t). Equations (29) and (32) generate a set of (2k-1M) equations in the unknown expansion coefficients {cnm:1≤n≤2k-1;0≤m≤M-1}. This system of equations can be solved with the aid of Newton’s iterative method (see [32]), to obtain the unknown components of the vector C, and hence the required spectral wavelets solution uk,M,λ(t) given by (25) can be obtained.
5. Convergence Analysis
In this section, we state and prove the following important theorem which ascertains that the ultraspherical wavelets expansion of a function f(x) with bounded second derivative converges uniformly to f(x).
Theorem 7.
A function f(t)∈Lω~2[0,1], ω~=(t-t2)λ-1/2, 0<λ<1, can be expanded as an infinite series of ultraspherical wavelets, which converges uniformly to f(x), provided |f′′(t)|⩽L. Explicitly, the expansion coefficients in (20) satisfy the inequality
(33)|cnm|<4L(1+λ)2(m+1+λ)2(m-2)4n5/2,∀n⩾1,m>2.
Proof.
From relations (16) and (19), one can write
(34)cnm=2k/2ξm,λ×∫(n-1)/2k-1n/2k-1f(t)Cm(λ)(2kt-2n-1)ω(2kt-n)dt,
and the integration of the right hand side of (34) by parts with the aid of relation (13) yields
(35)cnm=2(6-k)/2λξm,λm(m+2λ)×∫(n-1)/2k-1n/2k-1f′(t)Cm-1(λ+1)(2kt-2n-1)(2kt-n)hhhhhhhhhhhh×(1-2kt+n)ω(2kt-n)dt.
If we integrate (35) by parts again, and make use of the substitution: 2kt-2n-1=cosθ, then the coefficient cnm takes the form
(36)cnm=2(λ)2ξm,λ22λ+(5k-5)/2(m-1)2(m+2λ-1)2×∫0πf′′(1+2n+cosθ2k)Cm-2(λ+2)(cosθ)(sinθ)2λ+4dθ.
Now, assuming that m>2, taking into account the assumption |f′′(t)|⩽L, and with the aid of Theorem 4, we obtain
(37)|cnm|⩽2|(λ)2ξm,λ|22λ+(5k-5)/2(m-1)2(m+2λ-1)2×∫0π|f′′(1+2n+cosθ2k)||Cm-2(λ+2)(cosθ)|(sinθ)2λ+4dθ⩽2L|λ|(1+λ)|ξm,λ|22λ+(5k-5)/2(m-1)2(m+2λ-1)2×∫0π|Cm-2(λ+2)(cosθ)|(sinθ)2λ+4dθ<(2πL|λ|(1+λ)|ξm,λ|Γ(m+1+3λ2)Γ(3+λ2))×((2+λ2)23λ+(5k-5)/2(m-1)2(m+2λ-1)2hhhhhhh×Γ(λ+2)Γ(m+λ2)Γ(2+λ2))-1.
Knowing that λ>0, n<2k-1, and with the aid of relation (17), we get
(38)|cnm|<2L|λ|(1+λ)Γ((3+λ)/2)Γ(m+1+3λ/2)m!(m+λ)4λΓ(2+λ/2)Γ(m+λ/2)Γ(m+2λ)(m-2)4n5/2<4L(1+λ)2Γ(m+1+3λ/2)m!(m+λ)Γ(m+λ/2)Γ(m+2λ)(m-2)4n5/2<4L(1+λ)2(m+1+λ)2(m-2)4n5/2.
This completes the proof of the theorem.
Note. It is to be noted here that, for large values of m and n, and making use of the well-known Stirling’s formula (see, [33]), it can be easily shown that |cnm| is of O(n-5/2m-2).
6. Numerical Results and Comparisons
In this section, we apply the presented algorithms in Section 4, for obtaining numerical solutions for linear and nonlinear fractional-order Riccati differential equations. In addition, we compare our spectral ultraspherical algorithms with some other numerical algorithms that exist in the literature aiming to demonstrate the efficiency and rapid convergence of the proposed algorithms.
Example 8.
Consider the nonlinear Riccati differential equation (see, [34–36])
(39)Dαu(t)=1+2u(t)-u(t)2,t∈(0,1],u(0)=0,0<α≤1.
The exact solution for (39), when α=1, is given by
(40)u(t)=2tanh(2t-tanh-1(12))+1.
In Table 1, we compare the numerical solution resulting from the application of UWCM, for the case that corresponds to λ=k=1, M=10, with the generalized Euler method (GEM) and Laplace-Adomian-Padé method (LAPM) [34], while in Table 2, we compare the solution resulting from the application of UWCM, for the case that corresponds to λ=0, k=2, M=4, with the two solutions resulting from the application of homotopy perturbation method (HPM) developed in [35] and the variational iteration method (VIM) developed in [36]. In addition, Figure 1 illustrates different solutions for Example 8 in case of k=λ=α=1 for various values of M.
Comparison between UWCM, GEM, and LAPM for Example 8 for α=1.
t
GEM
LAPM
UWCM
Exact solution
LAPM error
UWCM error
0.1
0.1000000000
0.1102952044
0.1102951969
0.1102951969
7.5·10-9
1.4·10-12
0.2
0.2419000000
0.2419783394
0.2419767996
0.2419767996
1.5·10-6
2.7·10-12
0.3
0.3580039000
0.3951442714
0.3951048487
0.3951048487
3.9·10-5
4.9·10-12
0.4
0.5167880007
0.5682377001
0.5678121663
0.5678121663
4.2·10-4
6.8·10-12
0.5
0.6934386
0.7588607194
0.7580143934
0.7580143934
8.4·10-4
9.1·10-12
Comparison between UWCM, HPM, and VIM for Example 8 for α=1/2,3/4.
t
α=1/2
α=3/4
HPM
VIM
UWCM
HPM
VIM
UWCM
0.1
0.321730
0.592833
0.580926
0.216866
0.245446
0.244563
0.3
0.940941
1.174069
1.12057
0.654614
0.710050
0.710318
0.5
1.549439
1.473790
1.45668
1.132763
1.149016
1.15155
0.7
2.066523
1.646302
1.63391
1.594278
1.491949
1.49335
0.9
2.396839
1.756349
1.75008
1.962239
1.730575
1.73018
Different solutions of Example 8.
Example 9.
Consider the nonlinear Riccati differential equation (see, [36, 37])
(41)Dαu(t)=1-u(t)2,t∈(0,1],u(0)=0,0<α≤1.
The exact solution of (41), when α=1, is u(t)=tanht. In Table 3, the maximum absolute error E is listed for the case that corresponds to α=k=1, and for various values of M, while in Table 4, we give a comparison between the numerical solution resulting from the application of UWCM, for the case that corresponds to λ=1/2, k=2, M=5, with the two numerical solutions resulting from the application of the two methods, namely, variational iteration method (VIM) and second kind Chebyshev wavelets method (CUWM) developed, respectively, in [36, 37].
Maximum absolute error E for Example 9.
λ
M
E
M
E
M
E
M
E
M
E
1
5
2.76·10-6
8
3.72·10-8
11
5.17·10-11
14
1.44·10-13
17
2.22·10-16
1/2
3.77·10-6
5.09·10-9
1.32·10-11
3.26·10-14
2.22·10-16
0
3.03·10-6
1.28·10-8
2.90·10-11
7.24·10-14
4.44·10-16
-0.49
5.34·10-6
1.87·10-8
5.72·10-11
3.80·10-14
2.22·10-16
Comparison between UWCM, VIM, and CUWM for Example 9 for α=1/2,3/4.
t
α=1/2
α=3/4
VIM
CUWM
UWCM
VIM
CUWM
UWCM
0.2
0.436839
0.436737
0.430766
0.309976
0.309886
0.310516
0.4
0.553782
0.553802
0.556613
0.481632
0.481638
0.48269
0.6
0.621014
0.621026
0.645945
0.597783
0.597790
0.599666
0.8
0.666020
0.666016
0.666657
0.678850
0.678835
0.678914
Remark 10.
It is worth noting here that the results displayed in Table 3 show that the best absolute error resulting from the application of UWCM does not exceed (3×10-16), while the best error resulting from the application of the two methods, namely, VIM and CUWM developed in [36, 37], does not exceed (1.5×10-7). Moreover, the results presented in Table 4 appear to agree closely where the UWCM deviates somewhat, because of the expected high accuracy of the UWCM if compared with the results obtained by VIM and CUWM [36, 37].
Remark 11.
The results of Table 3 ascertain that the results that correspond to the first kind of Chebyshev wavelets expansion (in case of λ=0) are not always better than the other ultraspherical wavelets expansions (see, [17, 38]).
Example 12.
Consider the composite fractional oscillation equation (see, [36])
(42)D1/4u(t)+u(t)=t2+2Γ(11/4)t7/4,u(0)=0.
The exact solution for (42) is u(t)=t2. We apply the UWTM for problem (42), for the case that corresponds to k=1, M=3. In detail, the approximate solution u1,3,λ(t) is given by
(43)u1,3,λ(t)=c1,0ψ1,0(λ)(t)+c1,1ψ1,1(λ)(t)+c1,2ψ1,2(λ)(t)=c0C0(λ)(t)+c1C1(λ)(t)+c2C2(λ)(t)=c0+c1(2t-1)+c2(1+2λ-8(1+λ)(t-t2)1+2λ),
where, ci=c1,iξi, ξi is as given in (17), i=0,1,2.
In virtue of (27), the solution of the resulting linear system is
(44)c0=3+2λ8+8λ,c1=12,c2=1+2λ8+8λ,
and consequently
(45)u(t)=3+2λ8+8λ+12(2t-1)+1+2λ8+8λ(1+2λ-8(1+λ)(t-t2)1+2λ)=t2,
which is the exact solution.
7. Conclusions
This paper is concerned with presenting and implementing new algorithms for obtaining numerical spectral solutions for fractional-order Riccati differential equations. These methods are essentially based on employing the ultraspherical wavelets along with the application of the two spectral methods, namely, tau and collocation methods. The developed algorithms are applicable and easy in implementation and of high efficiency. The obtained numerical results indicate that the first kind of Chebyshev wavelets expansion is not always better than the other ultraspherical wavelets expansions.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Acknowledgment
This work was funded by the Deanship of Scientific Research (DSR), King Abdulaziz University, Jeddah, under Grant no. (130-269-D1435). The authors, therefore, acknowledge with thanks DSR technical and financial support.
Al-MdallalQ. M.SyamM. I.AnwarM. N.A collocation-shooting method for solving fractional boundary value problems201015123814382210.1016/j.cnsns.2010.01.020MR2652654ZBL1222.65078ÇenesizY.KeskinY.KurnazA.The solution of the Bagley-Torvik equation with the generalized Taylor collocation method2010347245246610.1016/j.jfranklin.2009.10.007MR2593729ZBL1188.65107Daftardar-GejjiV.JafariH.Adomian decomposition: a tool for solving a system of fractional differential equations2005301250851810.1016/j.jmaa.2004.07.039MR2105689ZBL1061.34003PodlubnyI.1999198New York, NY, USAAcademic PressMathematics in Science and EngineeringMR1658022JafariH.SeifiS.Solving a system of nonlinear fractional partial differential equations using homotopy analysis method20091451962196910.1016/j.cnsns.2008.06.019MR2474456ZBL1221.35439DasS.Analytical solution of a fractional diffusion equation by variational iteration method200957348348710.1016/j.camwa.2008.09.045MR2488620ZBL1165.35398MeerschaertM. M.TadjeranC.Finite difference approximations for two-sided space-fractional partial differential equations2006561809010.1016/j.apnum.2005.02.008MR2186432ZBL1086.65087BhrawyA. H.AlofiA. S.Ezz-EldienS. S.A quadrature tau method for fractional differential equations with variable coefficients201124122146215210.1016/j.aml.2011.06.016MR2826154ZBL1269.65068DohaE. H.BhrawyA. H.BaleanuD.Ezz-EldienS. S.On shifted Jacobi spectral approximations for solving fractional differential equations2013219158042805610.1016/j.amc.2013.01.051MR3037516Abd-ElhameedW. M.DohaE. H.YoussriY. H.Efficient spectral-Petrov-Galerkin methods for third- and fifth-order differential equations using general parameters generalized Jacobi polynomials2013361153810.2989/16073606.2013.779945MR3043668ZBL1274.65222DohaE. H.Abd-ElhameedW. M.YoussriY. H.Second kind Chebyshev operational matrix algorithm for solving differential equations of Lane-Emden type201323-24113117DohaE. H.Abd-ElhameedW. M.BassuonyM. A.New algorithms for solving high even-order differential equations using third and fourth Chebyshev-Galerkin methods201323656357910.1016/j.jcp.2012.11.009MR3020070CanutoC.HussainiM. Y.QuarteroniA.ZangT. A.1988New York, NY, USASpringerSpringer Series in Computational Physics10.1007/978-3-642-84108-8MR917480ElgindyK. T.Smith-MilesK. A.On the optimization of Gegenbauer operational matrix of integration2013393-451152410.1007/s10444-012-9289-5MR3116039ZBL1274.65078ElgindyK. T.Smith-MilesK. A.Solving boundary value problems, integral, and integro-differential equations using Gegenbauer integration matrices2013237130732510.1016/j.cam.2012.05.024MR2966907ZBL1259.65212DohaE. H.Abd-ElhameedW. M.Efficient spectral-Galerkin algorithms for direct solution of second-order equations using ultraspherical polynomials200224254857110.1137/S1064827500378933MR1951055ZBL1020.65088DohaE. H.Abd-ElhameedW. M.Accurate spectral solutions for the parabolic and elliptic partial differential equations by the ultraspherical tau method20051811244510.1016/j.cam.2004.11.015MR2145847ZBL1071.65136Abd-ElhameedW. M.DohaE. H.YoussriY. H.New wavelets collocation method for solving second-order multipoint boundary value problems using Chebyshev polynomials of third and fourth kinds201320139542839MR312151510.1155/2013/542839ZhuL.FanQ.Solving fractional nonlinear Fredholm integro-differential equations by the second kind Chebyshev wavelet20121762333234110.1016/j.cnsns.2011.10.014MR2877679ZBL06056883SadekI.AbualrubT.AbukhaledM.A computational method for solving optimal control of a system of parallel beams using Legendre wavelets2007459-101253126410.1016/j.mcm.2006.10.008MR2303403ZBL1117.49026ReidW. T.197286New York, NY, USAAcademic PressMathematics in Science and EngineeringMR0357936OldhamK. B.SpanierJ.1974New York, NY, USAAcademic PressMR0361633AndrewsG. E.AskeyR.RoyR.199971Cambridge, UKCambridge University PressEncyclopedia of Mathematics and its Applications10.1017/CBO9781107325937MR1688958GiordanoC.LaforgiaA.On the Bernstein-type inequalities for ultraspherical polynomials2003153243248RazzaghiM.YousefiS.Legendre wavelets direct method for variational problems200053318519210.1016/S0378-4754(00)00170-1MR1784947YousefiS. A.Legendre wavelets method for solving differential equations of Lane-Emden type200618121417142210.1016/j.amc.2006.02.031MR2270772ZBL1105.65080BabolianE.FattahzadehF.Numerical solution of differential equations by using Chebyshev wavelet operational matrix of integration2007188141742610.1016/j.amc.2006.10.008MR2327367ZBL1117.65178LiY.Solving a nonlinear fractional differential equation using Chebyshev wavelets20101592284229210.1016/j.cnsns.2009.09.020MR2602712ZBL1222.65087MaleknejadK.SohrabiS.RostamiY.Numerical solution of nonlinear Volterra integral equations of the second kind by using Chebyshev polynomials2007188112312810.1016/j.amc.2006.09.099MR2327098ZBL1114.65370Abd-ElhameedW. M.DohaE. H.YoussriY. H.New spectral second kind Chebyshev wavelets algorithm for solving linear and nonlinear second-order differential equations involving singular and Bratu type equations201320139715756MR313417910.1155/2013/715756BarrettR.BerryM.ChanT. F.1994Society for Industrial and Applied Mathematics (SIAM)10.1137/1.9781611971538MR1247007DemboR. S.EisenstatS. C.SteihaugT.Inexact Newton methods198219240040810.1137/0719025MR650059ZBL0478.65030LiY.-C.A note on an identity of the gamma function and Stirling's formula2007321267272MR2329236KhanN. A.AraA.KhanN. A.Fractional-order Riccati differential equation: analytical approximation and numerical results20132013article 185OdibatZ.MomaniS.Modified homotopy perturbation method: application to quadratic Riccati differential equation of fractional order200836116717410.1016/j.chaos.2006.06.041MR2372056ZBL1152.34311YinF.SongJ.LengJ. H.LuF.Couple of the variational iteration method and fractional-order legendre functions method for fractional differential equations20142014992876510.1155/2014/928765WangY.FanQ.The second kind Chebyshev wavelet method for solving fractional differential equations2012218178592860110.1016/j.amc.2012.02.022MR2921350ZBL1245.65090LightW. A.Are Chebysheve expansions really better198622180181