Research article

On spectral numerical method for variable-order partial differential equations

  • Received: 29 January 2022 Revised: 28 February 2022 Accepted: 07 March 2022 Published: 28 March 2022
  • MSC : 26A33, 35A02, 35A25

  • In this research article, we develop a powerful algorithm for numerical solutions to variable-order partial differential equations (PDEs). For the said method, we utilize properties of shifted Legendre polynomials to establish some operational matrices of variable-order differentiation and integration. With the help of the aforementioned operational matrices, we reduce the considered problem to a matrix type equation (equations). The resultant matrix equation is then solved by using computational software like Matlab to get the required numerical solution. Here it should be kept in mind that the proposed algorithm omits discretization and collocation which save much of time and memory. Further the numerical scheme based on operational matrices is one of the important procedure of spectral methods. The mentioned scheme is increasingly used for numerical analysis of various problems of differential as well as integral equations in previous many years. Pertinent examples are given to demonstrate the validity and efficiency of the method. Also some error analysis and comparison with traditional Haar wavelet collocations (HWCs) method is also provided to check the accuracy of the proposed scheme.

    Citation: Kamal Shah, Hafsa Naz, Muhammad Sarwar, Thabet Abdeljawad. On spectral numerical method for variable-order partial differential equations[J]. AIMS Mathematics, 2022, 7(6): 10422-10438. doi: 10.3934/math.2022581

    Related Papers:

    [1] Obaid Algahtani, M. A. Abdelkawy, António M. Lopes . A pseudo-spectral scheme for variable order fractional stochastic Volterra integro-differential equations. AIMS Mathematics, 2022, 7(8): 15453-15470. doi: 10.3934/math.2022846
    [2] Zahra Pirouzeh, Mohammad Hadi Noori Skandari, Kamele Nassiri Pirbazari, Stanford Shateyi . A pseudo-spectral approach for optimal control problems of variable-order fractional integro-differential equations. AIMS Mathematics, 2024, 9(9): 23692-23710. doi: 10.3934/math.20241151
    [3] Bin Fan . Efficient numerical method for multi-term time-fractional diffusion equations with Caputo-Fabrizio derivatives. AIMS Mathematics, 2024, 9(3): 7293-7320. doi: 10.3934/math.2024354
    [4] Amjid Ali, Teruya Minamoto, Rasool Shah, Kamsing Nonlaopon . A novel numerical method for solution of fractional partial differential equations involving the $ \psi $-Caputo fractional derivative. AIMS Mathematics, 2023, 8(1): 2137-2153. doi: 10.3934/math.2023110
    [5] Imran Talib, Md. Nur Alam, Dumitru Baleanu, Danish Zaidi, Ammarah Marriyam . A new integral operational matrix with applications to multi-order fractional differential equations. AIMS Mathematics, 2021, 6(8): 8742-8771. doi: 10.3934/math.2021508
    [6] Jeong-Kweon Seo, Byeong-Chun Shin . Reduced-order modeling using the frequency-domain method for parabolic partial differential equations. AIMS Mathematics, 2023, 8(7): 15255-15268. doi: 10.3934/math.2023779
    [7] Xiaojun Zhou, Yue Dai . A spectral collocation method for the coupled system of nonlinear fractional differential equations. AIMS Mathematics, 2022, 7(4): 5670-5689. doi: 10.3934/math.2022314
    [8] Khalid K. Ali, Mohamed A. Abd El Salam, Mohamed S. Mohamed . Chebyshev fifth-kind series approximation for generalized space fractional partial differential equations. AIMS Mathematics, 2022, 7(5): 7759-7780. doi: 10.3934/math.2022436
    [9] A. S. Mohamed . Fibonacci collocation pseudo-spectral method of variable-order space-fractional diffusion equations with error analysis. AIMS Mathematics, 2022, 7(8): 14323-14337. doi: 10.3934/math.2022789
    [10] Samia Bushnaq, Kamal Shah, Sana Tahir, Khursheed J. Ansari, Muhammad Sarwar, Thabet Abdeljawad . Computation of numerical solutions to variable order fractional differential equations by using non-orthogonal basis. AIMS Mathematics, 2022, 7(6): 10917-10938. doi: 10.3934/math.2022610
  • In this research article, we develop a powerful algorithm for numerical solutions to variable-order partial differential equations (PDEs). For the said method, we utilize properties of shifted Legendre polynomials to establish some operational matrices of variable-order differentiation and integration. With the help of the aforementioned operational matrices, we reduce the considered problem to a matrix type equation (equations). The resultant matrix equation is then solved by using computational software like Matlab to get the required numerical solution. Here it should be kept in mind that the proposed algorithm omits discretization and collocation which save much of time and memory. Further the numerical scheme based on operational matrices is one of the important procedure of spectral methods. The mentioned scheme is increasingly used for numerical analysis of various problems of differential as well as integral equations in previous many years. Pertinent examples are given to demonstrate the validity and efficiency of the method. Also some error analysis and comparison with traditional Haar wavelet collocations (HWCs) method is also provided to check the accuracy of the proposed scheme.



    Fractional calculus has got great popularity among the researchers of different areas of science and engineering. It has been reported that dynamical systems typically undergo two stages of development, one is from integer-order dynamical systems to fractional-order systems and techniques in the domains of solid mechanics [1], physics [2], finance [3], population growth [4], physiology [5], electro-mechanical [6], statistical mechanics [7], systems of fractional differential equations(FDEs) of different kinds appear frequently. Many engineering, physical and biological problems can be modeled using the applications of fractional calculus for more better results than classical calculus. The results of FDEs and integral equations are more accurate and precise as compared to conventional differential equations, see examples [8,9,10,11,12]. On the basis of uniqueness and existence, plenty of research articles have been published (we refer [13,14,15,16,17]).

    Here we state that the area where fractional derivative of constant real or complex order has been explored very well. Further huge amount of research work in the said area have been published. As in fractional calculus, the order of differentiation and integration is arbitrary or it can also be a function. Hence the generalization of Reimann-Liouville fractional order to variable-order was founded by Ross and Samko in 1993. Thus Ross and Samko gave the spearheading and initiative work on variable-order operator of differentiation and integration (see [18]). A while later, several contributions have been reported including applications of variable-order operators in mathematical models of various phenomenons (we refer [19,20,21]). Consequent studies explored that variable-order fractional calculus can be exceptionally valuable in areas like viscous flows, mechanics and modeling many phenomenons. In this regards various articles have been written to investigate the aforesaid area from different aspects including existence theory, numerical and analytical results (see [22,23,24,25]). Due to these facts said derivatives have the ability to describe real problems in more comprehensive ways. To the best of our knowledge the numerical results based on operational matrices for variable order FDEs have very rarely investigated.

    The theory of numerical approximation for solving differential equations of variable order is getting to be increasingly imperative. Since spectral methods based on polynomial basis have been considered in previous time significantly for the computation of numerical results of various problems. The aforesaid methods are stable and powerful as compared to other numerical methods like finite element and residual power series methods, radial base function methods, etc. Also the mentioned methods applied to any linear problems are stable and having faster convergence rate (see some detail in [26]). Here we remark that spectral method based on polynomials basis including Legendre, Jacobi and Bernstein polynomials, etc have been reported stable apply to any linear problems (see detail in [27]). Further the said methods have been considered very well due to the following points of interest like

    ● simple to utilize and having clear and evident form;

    ● offer a suitable framework to approximate the solution of a problem;

    ● avoiding complex calculations due to simpler weight functions;

    ● spectral methods use the idea of global representations to find greater order approximations;

    ● can be easily applied to any differential equations of variable and constant order.

    ● as compared to finite difference methods, spectral methods are global techniques.

    Here it is interesting to say that there is significant difference between difference and spectral methods. The spectral methods utilize basis functions which are nonzero over the entire domain, while difference methods use basis functions that are nonzero only on very small sub domains. So far we know the aforesaid methods have been applied for PDEs with constant real or complex order. But never use in computation of numerical solutions of variable-order PDEs.

    Motivated and inspired from the mentioned work above, the objective of this manuscript is devoted to form a scheme for the computation of numerical solutions to a class of variable-order PDEs. To achieve this goal, we utilize operational matrices of integration and differentiation based on shifted Legendre polynomials. We extend operational matrices of differentiation and integration using shifted Legendre polynomials from simple real or complex order to any variable -order. Hence an algorithm is established for a class of multi-terms variable- order PDEs to compute numerical solutions. These operational matrices are based on shifted Legendre polynomials of two variables. Based on these matrices, we reduce the proposed variable order PDE to some algebraic equation. Keeping in mind, that the proposed method omits discretization and collocation which save much more time and memory. We investigate the class of linear multi terms variable order PDE studied earlier in [28] under variable order derivative as

    c1α(t)U(t,x)tα(t)+c2α(t)U(t,x)tα(t)2xα(t)2+c3α(t)U(t,x)xα(t)+c4β(t)U(t,x)tβ(t)+c5β(t)U(t,x)xβ(t)+c6U(t,x)=g(t,x), (1.1)

    with initial conditions

    U(0,x)=θ(x),  Ut(0,x)=ϕ(x), (1.2)

    where c1,c2,c3,c4,c5 and c6 are constants

    U(t,x), g(t,x)C([0,1]×[0,1])  and α(t)(1,2], β(t)(0,1].

    Also

    α, βC[0,1]

    If we select α(t)=2 and β(t)=1 in (1.1), it becomes integer order PDEs. Further, the problem (1.1) represents various classes of PDEs, by selecting difference values for coefficient and taking α(t)=2 and β(t)=1 as:

    ● If (c2)24(c1)(c2)>0, then problem (1.1) becomes a class of hyperbolic PDEs.

    ● If (c2)24(c1)(c2)=0, then problem (1.1) becomes a class of parabolic PDEs.

    ● If (c2)24(c1)(c2)0, then the problem (1.1) becomes a class of elliptic PDEs.

    ● If c1=1, c2=0, c3=c2, c4=c5=c6=0, g(t,x)=x, the problem (1.1) becomes the famous wave PDE with source term in space variable been studied in [35] by using HWCs method.

    In same way fixing the coefficients and orders, the famous Poisson PDE and Laplace PDE become special cases of the problem (1.1). The mentioned PDEs have wide range applications in the field of mechanics and electro-magnetics, solitons and turbulent flow theory. We utilize our adopted procedure to present the numerical results graphically. Also comparison with HWCs method is given. Further, absolute errors are recorded accordingly. For some recent work in numerical analysis of various problems, we refer [31,33] and the references therein.

    This article is organized as: In Section 2, basics results are recalled from literature of fractional calculus. Section 3 is devoted to the development of operational matrices for variable order integration and differentiation. Section 4 is devoted to the establishment of the proposed method for variable-order PDEs and also supporting examples along with the graphs are provided. Last section is devoted to brief conclusion.

    In this section, we recall some basic results which can be found in [18,29,30].

    Definition 2.1. Let α>0 be continues and bounded function, then the variable- order Riemann–Liouville integral of a function hL[0,1] is defined by

    0Iα(t)th(t)=1Γ(α(t))t0(tς)α(ς)1h(ς)dς, t>0,

    provided that integral exists on right side.

    Definition 2.2. The Caputo derivative of a function hC[0,1] with variable- order n1<α(t)n can be defined as

    CDα(t)h(t)=1(Γnα(t))t0(tς)nα(ς)1h(n)(ς)dς. (2.1)

    From (2.1), the following results hold

    C0Dα(t)ttm={0,m<α(t),Γ(m+1)Γ(m+1α(t))tmα(t),otherwise, (2.2)

    Definition 2.3. [32] The Legendre polynomials on the interval [1,1] are recalled as below

    Łr+1(x)=2r+1r+1xŁrxrr+1Łr1(x), (2.3)

    where Ł0(x)=1 and Łx(x)=x.

    By the use of transformation x=2t+1L1, we obtain the recursive relation for shifted Legendre polynomials denoted by ŁL,r(t) on t[0,L] as

    ŁL,r+1(t)=2r+1r+1(2tL1)ŁL,r(t)rr+1ŁL,r+1(t), r=1,2,3,. (2.4)

    Further the shifted Legendre polynomial of degree n in [0,1] is expressed as

    Łn(t)=nr=0(1)n+r[Γ(n+r+1)Γ(nr+1)Γ2(r+1)]tr,  where  r=0,1,2,...,n.

    Further, the orthogonality condition for shifted Legendre polynomials over [0, 1] is given by

    10ŁL,q(t)ŁL,r(t)ds={12r+1, r=q,0, otherwise. (2.5)

    Further in two variables, the said polynomials are expressed as

    Łn(x,t)=Łr(x)Łq(t),  n=M+1, r,q=0,1,2,...,n.

    The orthogonality is computed as

    1010Łr(x)Łq(t)¯Łr(x)¯Łq(t)dxdt={1(2r+1)(2q+1), r=q,0, otherwise.

    Let B=C([0,1]×[0,1]) be the Banach space of square integrable functions. Then any function UB can be approximated in terms of two variables shifted Legendre polynomials as

    U(x,t)nr=0nq=0Cr,qŁr(x)Łq(t), (2.6)

    where

    Cr,q=(2r+1)(2q+1)1010U(x,t)Łr(x)Łq(t)dxdt.

    On using Cr,q=Cn, then in analytical form of the said polynomial (2.6) can be written in matrix form as

    ΥL,M(t)M2n=1CnŁr(x,t)=XMϕM2(x,t),

    where XM is the coefficient vector of dimension 1×M2 and ϕM2(x,t) function vector of dimension M2×1.

    Here we can prove the convergence of the proposed method in same line as done in [36].

    Theorem 2.4. Suppose that U(t,x) and Un(t,x) are the exact and the approximated solution respectively computed by the proposed algorithm, then

    Un(t,x)U(x,t)14(1n)n+1[Ω1+Ω2+Ω34(1n)n+1], (2.7)

    where Ω1,Ω2,Ω3 are terminated constants given by

    Ω1=max(x,t)X|n+1U(x,t)xn+1|,Ω2=max(x,t)X|n+1U(x,t)tn+1|,Ω3=max(x,t)X|2n+2U(x,t)xn+1tn+1|. (2.8)

    Proof. Let Un(x,t) is the approximate solution computed in terms of interpolated polynomial for U(x,t) at point (xi,ti) by means of aforesaid method such that xi=in and ti=in. Then using (2.8) and after computation, briefly we can write the error bounds as

    Un(x,t)U(x,t)14(1n)n+1[Ω1+Ω2+Ω34(1n)n+1].

    Therefore

    Un(x,t)U(x,t)14(1n)n+1[Ω1+Ω2+Ω34(1n)n+1]. (2.9)

    Now from (2.9), the error bound for the maximum absolute value

    14(1n)n+1[Ω1+Ω2+Ω34(1n)n+1].

    Since the space B bounds and the absolute error bound coincides, Therefore when n, the right side of (2.9) goes to zero. Hence one has

    Un(x,t)U(x,t)0,  for all  (x,t)B  as  n.

    This show the convergence of the procedure.

    This section is devoted to the construction of operational matrices of variable- order differentiation and integration based on shifted Legendre polynomials. For the required purpose we follow the same rules as followed in [34].

    Theorem 3.1. If ϕM2(x,t) be the function vector, thenvariable- order integration with respect to t of ϕM2(x,t) yields

    0Iα(t)t[ϕM2(x,t)]P(α(t),x)M2×M2ϕM2(x,t), (3.1)

    where

    Pα(t),xM2×M2=[ζ1,1,zζ1,2,zζ1,r,zζ1,M2,zζ2,1,zζ2,2,zζ2,r,zζ2,M2,zζv,1,zζv,2,zζv,r,zζv,M2,zζM2,1,zζM2,2,zζM2,r,zζM2,M2,z] (3.2)

    where r=Mi+j+1, v=Ma+b+1, ζv,r,z=σi,j,b,z for i,j,a,b=0,1,2,,m

    σi,j,b,z=an=0a,z,α(t)i,j,b.
    i,j,b=ςj,bip=0(1)ip(2i+1)Γ(n+p+1)Γ(p+1)Γ(p+1)Γ(n+α(t)+p+2) (3.3)

    and

    a,z,α(t)=(1)azΓ(a+z+1)Γ(az+1)Γ(z+1)Γ(1+z+α(t)). (3.4)

    Theorem 3.2. The variable-order differentiation of ϕM2(x,t) with respect to t is given as

    α(t)tα(t)[ϕM2(x,t)]Rα(t),tM2×M2ϕM2(x,t), (3.5)

    where Rα(t),tM2×M2 is the operationalmatrix of differentiation and is given as

    Rα(t),tM2×M2=[Φ1,1,zΦ1,2,zΦ1,r,zΦ1,M2,zΦ2,1,zΦ2,2,zΦ2,r,zΦ2,M2,zΦv,1,zΦv,2,zΦv,r,zΦv,M2,zΦM2,1,zΦM2,2,zΦM2,r,zΦM2,M2,z] (3.6)

    r=Mi+j+1, v=Ma+b+1, Φv,r,z=Πi,j,a,b,z, for i,j,b,z=0,1,2,,n

    Πi,j,a,b,z=az=0Λa,z,α(t)Λi,j,b, (3.7)
    Λi,j,b=ςj,bip=0(1)i1(2i+1)Γ(i+1)Γ(z+p+1)Γ(p+1)Γ(p+1)Γ(zα(t)+p+2) (3.8)

    and

    Λa,z,α(t)=(1)azΓ(a+z+1)Γ(az+1)Γ(z+1)Γ(1+zα(t)). (3.9)

    Theorem 3.3. The fractional order derivative of ϕM2(x,t) w.r.t x and t is given by

    α(t)tα(t)2xα(t)2ϕM2(x,t)R(α(t),x,t)M2×M2ϕM2(x,t) (3.10)

    and

    Rα(t),x,tM2×M2=[Ω1,1,zΩ1,2,zΩ1,r,zΩ1,M2,zΩ2,1,zΩ2,2,zΩ2,r,zΩ2,M2,zΩv,1,zΩv,2,zΩv,r,zΩv,M2,zΩM2,1,zΩM2,2,zΩM2,r,zΩM2,M2,z], (3.11)

    where r=Mi+j+1, v=Ma+b+1, Ωv,r,z=i,j,a,b, for ,a,b,i,j=0,1,2,,k

    i,j,a,b=az=0ϖa,z,α(t)ϖi,j,b (3.12)
    ϖi,j,b=ςj,bip=0(1)il(2i+1)Γ(i+p+1)Γ(zα(t)+p+1)Γ(p+1)Γ(p+1)Γ(ip+1)Γ(zα(t)+p+2)Lα(t) (3.13)

    and

    ϖa,z,α(t)=(1)azΓ(a+z+1)Γ(az+1)Γ(z+1)Γ(1+zα(t))Lz. (3.14)

    Remark 1. Operational matrices based on shifted Legendre polynomials play important roles in numerical analysis.

    On the bases of operational matrices developed previously, our considered problem (1.1) is reduced into simple Sylvester type equation. Then the algebraic equation is solved for the unknown coefficient matrix to compute the required numerical result for U(x,t). Consider the approximation as

    α(t)U(t,x)tα(t)=HTM2ϕM2(t,x). (4.1)

    By using 3.1), one has

    U(t,x)l1l2t=HTM20Iα(t)[ϕM2(t,x)]. (4.2)

    Now to find the values of l1 and l2, we use initial conditions U(0,x)=θ(x) and Ut(0,x)=ϕ(x) in (4.2). Hence one has l1=θ(x) and l2=ϕ(x). By inserting these values in (1.1), one has

    U(t,x)=HTM2P(α(t),t)M2×M2ϕM2(t,x)+θ(x)+tϕ(x).

    We can write as

    U(t,x)=HTM2P(α(t),t)M2×M2ϕM2(t,x)+GM2ϕM2(t,x),

    where

    θ(x)+tϕ(x)=GM2ϕM2(t,x)

    and

    U(t,x)=[HTM2Pα(t),tM2×M2+GM2]ϕM2(t,x). (4.3)

    Now, from (4.3), we calculate the approximation of remaining terms of (1.1) as

    α(t)U(t,x)xα(t)=[HTM2Pα(t),tM2×M2+GM2]Rα(t),xM2×M2ϕM2(t,x)β(t)U(t,x)tβ(t)=[HTM2Pα(t),xM2×M2+GM2]Rβ(t),tM2×M2ϕM2(t,x)β(t)U(t,x)xβ(t)=[HTM2Pβ(t),xM2×M2+GM2]Rβ(t),xM2×M2ϕM2(t,x) (4.4)
    α(t)U(t,x)tα(t)2xα(t)2=[HTM2Pα(t),tM2×M2+GM2]Rα(t),t,xM2×M2ϕM2(t,x)g(t,x)=QM2ϕM2(t,x). (4.5)

    So our considered class of variable -order PDEs (1.1) by using (4.3) and (4.4) reduces to the following form

    c1HTM2ϕ2M(t,x)+c2[HTM2Pα(t),tM2×M2+G2M]R(α(t),t,x)M2×M2ϕ2M(t,x)+c3[HTM2Pα(t),tM2×M2+G2M]R(α(t),x)M2×M2ϕ2M(t,x)+c4[HTM2Pα(t),tM2×M2+G2M]R(β(t),t)M2×M2ϕ2M(t,x)+c5[HTM2Pα(t),tM2×M2+G2M]R(β(t),x)M2×M2ϕ2M(t,x)+c6[HTM2Pα(t),tM2×M2+G2M]ϕ2M(t,x)=QM2ϕ2M(t,x). (4.6)

    By further simplification, (4.6) yields

    c1HTM2+HM2P(α(t),t)M2×M2[c2R(α(t),t,x)M2×M2+c3Rα(t),xM2×M2+c4R(β(t),t)M2×M2+c5R(β(t),x)M2×M2+c6IM2×M2]+GM2×M2[c2R(α(t),t,x)M2×M2+c3R(α(t),x)M2×M2+c4R(β(t),t)M2×M2+c5R(β(t),x)M2×M2+c6R(α(t),t,x)M2×M2]QM2=OM2. (4.7)

    Further simplification of (4.7) yields

    A=P(α(t),t)M2×M2[c2Rα(t),t,xM2×M2+c3Rα(t),xM2×M2+c4Rβ(t),tM2×M2+c5Rβ(t),xM2×M2+c6IM2×M2]. (4.8)

    Using B=GM2AQM2 and (4.8) in (4.7), one has

    c1HTM2+HM2TA+B=0. (4.9)

    Hence (4.9) is of the form X+XA+B=0. Which can be solved using Matlab by Gauss elimination method to get X for the required numerical solution.

    The following examples are provided to demonstrate the method.

    Example 1.

    {α(t)U(t,x)tα(t)9α(t)U(t,x)xα(t)4β(t)U(t,x)xβ(t)9U(t,x)=12exp(t)sin(x)13exp(t)sin(t),  1<α(t)2, 0<β(t)1,U(0,x)=sin(x), Ut(0,x)=sin(x) (5.1)

    At α(t)=2, β(t)=1, the exact solution is given by

    U(t,x)=exp(t)sin(x).

    Now we are approximating the solution through the suggested method and also represented graphically in Figure 1 respectively. Now in Table 1, we give absolute errors at α(t)=1+exp(t), β=1

    Figure 1.  Graphical presentation of approximate solutions and absolute error of Example 1 at α(t)=1+exp(t), β(t)=1 and Scale level 6.
    Table 1.  Maximum absolute error at various values of (x,t) and scale level for Example 1.
    (x,t) Scale level Absolute error
    (0.1,0.1) 6 2×102
    (0.1,0.2) 8 2.5×102
    (0.2,0.3) 10 2×103
    (0.3,0.4) 12 4×104
    (0.5,0.5) 14 5×105
    (0.6,0.7) 16 6×106
    (0.9,0.9) 14 4×106
    (1.0,1.0) 16 7×107

     | Show Table
    DownLoad: CSV

    Example 2.

    {α(t)U(t,x)tα(t)2α(t)U(t,x)tα(t)2xα(t)2+α(t)U(t,x)xα(t)+4β(t)U(t,x)tβ(t)+4β(t)U(t,x)xβ(t)9U(t,x)=g(t,x),   1<α(t)2,1<β(t)2,U(0,x)=sin(x),  Ut(0,x)=cos(x). (5.2)

    For α(t)=2 and β(t)=1, the integer order solution is given by U(t,x)=sin(x+t) and g(t,x)=8cos(x+t)9sin(x+t). Now we approximate the solution through the suggested method and presents graphically in Figure 3 respectively. Now in Table 2, we give absolute errors at various fractional order.

    Figure 2.  Graphical presentation of approximate solutions and absolute error of Example 1 at α(t)=2, β(t)=1exp(t) and Scale level 6.
    Figure 3.  Graphical presentation of approximate solutions and absolute error of Example 2 at various values of α=2sin(t), β(t)=1exp(t), and scale level 6.
    Table 2.  Maximum absolute error at various values of (x,t) and scale level for Example 2.
    (x,t) Scale level Absolute error
    (0.1,0.1) 6 1.0×102
    (0.1,0.2) 8 2.0×102
    (0.2,0.3) 10 3.0×103
    (0.3,0.4) 12 4.0×104
    (0.5,0.5) 14 3.9×104
    (0.6,0.7) 16 4.0×105
    (0.9,0.9) 14 5.0×106
    (1.0,1.0) 16 7.0×107

     | Show Table
    DownLoad: CSV

    Example 3. Let choose the values of constants in (1.1) such that c1=1, c2=0, c3=c2, c4=c5=c6=0 and g(t,x)=x, we get the following wave equation with source term or in-homogeneous differential equation [35]

    α(t)U(x,t)αttc2βtU(x,t)βx=x, 0<x<1, t>0, (5.3)

    with

    U(x,0)=0,  Ut(x,0)=0,  0<x<1 (5.4)

    and

    U(0,t)=0,  U(1,t)=0,  t>0. (5.5)

    For α=β=2 (3) gives exact solution of the following form

    U(x,t)=2π3c2k=1(1)k1k3[1cos(kπct)]sin(kπx). (5.6)

    Now in Table 3, we give absolute errors at various fractional order. Further in Table 4, we compared the maximum absolute error at various fractional order of our solution and compared with HWCs method [35]. Let the maximum absolute error be UˉU, then Also we compared the exact solution with the given approximate solution in Figures 5 and (6) respectively. We have compared our results with the results of HWCs method for the Example 3 by taking scale level 10 and for HWCs method taking collocation points 32. We see that the adopted spectral method is much more better than the HWCs method. Here if we increase the scale level further, the accuracy can be further enhanced.

    Table 3.  Maximum absolute error at various values of (x,t) and scale level for Example 2, at α=2,β=1.
    (x,t) Scale level Absolute error
    (0.1,0.1) 6 2.9×103
    (0.1,0.2) 8 2.5×103
    (0.2,0.3) 10 3.8×104
    (0.3,0.4) 12 3.4×104
    (0.5,0.5) 14 4.8×105
    (0.6,0.7) 16 4.5×105
    (0.9,0.9) 14 5.6×107
    (1.0,1.0) 16 1.9×1010

     | Show Table
    DownLoad: CSV
    Table 4.  Maximum absolute error at various values of (x,t) and scale level M=10 for Example 3 at α=2,β=1.
    (x,t) Value of UˉU at given method for M=10 UˉU in [35] M=32
    (0.1,0.1) 5.56×106 9.9×106
    (0.1,0.2) 4.78×106 2.9×105
    (0.2,0.3) 2.9×107 4.9×105
    (0.3,0.4) 3.9×108 6.7×105
    (0.5,0.5) 4.7×108 1.9×104
    (0.6,0.7) 6.6×109 2.0×104
    (0.9,0.9) 5.3×1010 2.96×104
    (1.0,1.0) 8.6×1012 1.9×105

     | Show Table
    DownLoad: CSV
    Figure 4.  Graphical presentation of approximate solutions and absolute error of Example 2 at various values of α=t2+1.52, with 0<t<≤1,β(t)=1t22, and scale level 6.
    Figure 5.  Graphical presentation of approximate solutions and absolute error of Example 3 at various values of α=1+exp(πt), β(t)=1, and scale level 6.
    Figure 6.  Graphical presentation of approximate solutions and absolute error of Example 3 at various values of α=2cos(t), β(t)=1, and scale level 6.

    We have presented an algorithm for the numerical solution of variable- order PDEs by using the concept of operational matrices. The concerned matrices have been created by using shifted Legendre polynomials under variable -order Caputo derivative and Riemann-Liouville integration. The aforesaid matrices have the ability to reduce the proposed problem to some algebraic equation. With the help of Matlab, we have evaluated the mentioned equation to compute numerical results. Three test problems have been investigated by using different kinds of variable orders. We see that the proposed method provides more better results. The concerned differential and integral operators have greatest degree of freedom as compared to classical fractional order. We have also used different spaces points and computed the maximum absolute error for various scale levels. By increasing the scale level, the efficiency of the method can be enhanced and the accuracy be improved further. Further, we have compared our results in Example 3 with the results of HWCs. We see that spectral method adopted in this work provides much more better results than the aforementioned numerical method. Further the procedure can be extended to other nonlinear problems also. Further variable -order PDEs can also be used as power full tools to investigate various real world problems like diffusion, blood flow, etc. For the mentioned problems such type numerical methods are more powerful and helpful.

    Authors Kamal Shah and Thabet Abdeljawad would like to thank Prince Sultan University for paying the APC and support through the research lab TAS.

    There is no competing interest regarding this work.



    [1] Y. A. Rossikhin, M. V. Shitikova, Application of fractional calculus for dynamic problems of solid mechanics: novel trends and recent results, ASME. Appl. Mech. Rev., 63 (2010), 010801. https://doi.org/10.1115/1.4000563 doi: 10.1115/1.4000563
    [2] R. Hilfer, Applications of fractional calculus in physics, World Scientific, Singapor, 2000. https://doi.org/10.1142/3779
    [3] V. E. Tarasov, On history of mathematical economics: Application of fractional calculus, Mathematics, 7 (2019), 509. https://doi.org/10.3390/math7060509 doi: 10.3390/math7060509
    [4] J. V. D. C. Sousa, E.C. de Oliveira, On two new operators in fractional calculus and application, arXiv preprint arXiv, 2017.
    [5] B. J. West, Fractal physiology and the fractional calculus: a perspective, Front. physiol, 1 (2010), 12.
    [6] M. P. Aghababa, Chaos in a fractional-order micro-electro-mechanical resonator and its suppression, Chinese Phys. B, 21 (2012), 100505. https://doi.org/10.1088/1674-1056/21/10/100505 doi: 10.1088/1674-1056/21/10/100505
    [7] A. Carpinteri, F. Mainardi, Fractals and Fractional Calculus in Continuum Mechanics, Springer, Verlage, 2014.
    [8] R. Gorenflo, F. Mainardi, E. Scalas, M. Raberto, Fractional calculus and continuous time finance III, The diffusion limit, Mathe. Finance, Birkhäuser, Basel, 2001 (2001), 171–180.
    [9] F. Liu, V. Anh, I. Turner, Numerical solution of the space fractional Fokker–Planck equation, J. Comp. Appl. Math., 66 (2005), 209–219.
    [10] R. Metzler, J. Klafter, The random walks guide to anomalous diffusion: A fractional dynamics approach, Phys. Reports, 339 (2000), 1–77. https://doi.org/10.1016/S0370-1573(00)00070-3 doi: 10.1016/S0370-1573(00)00070-3
    [11] R. Metzler, J. Klafter, Boundary value problems for fractional diffusion equations, Phys. A: Stat. Mech. Appl., 278 (2005), 107–125.
    [12] A. Saadatmandi, M. Deghan, Numerical solution of the a mathematical formation in tumer angionesis via tau method, Commun. Numer. Methods Eng., 24 (2008), 1467–1474. https://doi.org/10.1002/cnm.1045 doi: 10.1002/cnm.1045
    [13] R.A. Khan, M. Rehman, Existence of multiple positive solutions for a general system of fractional differential equations, Commun. Appl. Nonlinear Anal., 18 (2011), 25–35.
    [14] K. I. Isife, Positive solutions of a class of nonlinear boundary value fractional differential equations, J. Frac. Calc. Nonl. Sys., 2 (2021), 12–30.
    [15] M. Rehman, R.A. Khan, Existence and uniqueness of solutions for multi-point boundary value problems for fractional differential equations, Appl. Math. Lett. 23 (2010), 1038–1044. https://doi.org/10.1016/j.aml.2010.04.033
    [16] E. Ziada, Numerical solution for multi-term fractional delay differential equations, J. Frac. Calc. Nonl. Sys., 2 (2021), 1–12.
    [17] M. Rehman, R. A. Khan, J. Henderson, Existence and uniqueness of solutions for nonlinear fractional differential equations with integral boundary conditions, Fract. Dyn. Syst., 1 (2011), 29–43. https://doi.org/10.7153/fdc-01-02 doi: 10.7153/fdc-01-02
    [18] S. G. Samko, B. Ross, Integration and differentiation to a variable fractional order, Integral Transforms Spec. Funct., 1 (1993), 277–300.
    [19] H. Dehestani, Y. Ordokhani, M. Razzaghi, Numerical solution of variable-order time fractional weakly singular partial integro-differential equations with error estimation, Mathematical Modelling and Analysis, 25 (2020), 680–701. https://doi.org/10.3846/mma.2020.11692 doi: 10.3846/mma.2020.11692
    [20] H. Dehestani, Y. Ordokhani, M. Razzaghi, A novel direct method based on the Lucas multiwavelet functions for variable-order fractional reaction-diffusion and subdiffusion equations, Numer. Linear Algebra Appl., 28 (2021), e2346. https://doi.org/10.1002/nla.2346 doi: 10.1002/nla.2346
    [21] H. Dehestani, Y. Ordokhani, M. Razzaghi, Modified wavelet method for solving multitype variable-order fractional partial differential equations generated from the modeling of phenomena, Math. Sci., (2021), 1–17.
    [22] J. Jiang, J. L. Guirao, T. Saeed, The existence of the extremal solution for the boundary value problems of variable fractional order differential equation with causal operator, Fractals, 28 (2020), 2040025. https://doi.org/10.1142/S0218348X20400253 doi: 10.1142/S0218348X20400253
    [23] S. Sarwar, On the existence and stability of variable order caputo type fractional differential equations, Fractal Fract., 6 (2022), 51. https://doi.org/10.3390/fractalfract6020051 doi: 10.3390/fractalfract6020051
    [24] C. F. M. Coimbra, Mechanics with variable-order differential operators, Ann. Phys., 12 (2003), 692–703. https://doi.org/10.1002/andp.200310032 doi: 10.1002/andp.200310032
    [25] P. Zhuang, F. Liu, V. Anh, I. Turner, Numerical methods for the variable-order fractional advection-diffusion equation with a nonlinear source term, SIAM J. Numer. Anal., 47 (2009), 1760–1781. https://doi.org/10.1137/080730597 doi: 10.1137/080730597
    [26] C. Canuto, M. Y. Hussaini, A. Quarteroni, T. A. Zang, Spectral Methods: Fundamentals in Single Domains, Springer Science and Business Media, Berlin, 2007.
    [27] D. Gottlieb, L. Lustman, E.Tadmor, Stability analysis of spectral methods for hyperbolic initialboundary value systems, SIAM J. Numer. Anal., 24 (1987), 241–256. https://doi.org/10.1137/0724020 doi: 10.1137/0724020
    [28] K. Shah, M. Akram, Numerical treatment of non-integer order partial differential equations by omitting discretization of data, Comput. Appl. Math., 37 (2018), 6700–6718. https://doi.org/10.1007/s40314-018-0706-3 doi: 10.1007/s40314-018-0706-3
    [29] Y. Chen, L. Liu, B. Li, Y. Sun, Numerical solution for the variable order linear cable equation with Bernstein polynomials, Appl. Math. Comput., 238 (2014), 329–341. https://doi.org/10.1016/j.amc.2014.03.066 doi: 10.1016/j.amc.2014.03.066
    [30] D. Valério, J. S. Da Costa, Variable-order fractional derivatives and their numerical approximations, Signal process., 91 (2011), 470–483. https://doi.org/10.1016/j.sigpro.2010.04.006 doi: 10.1016/j.sigpro.2010.04.006
    [31] Y. Chen, Y. Wei, D. Liu, H. Yu, Numerical solution for a class of nonlinear variable order fractional differential equations with Legendre wavelets, Appl. Math. Lett., 46 (2015), 83–88.
    [32] M. A. Zaky, D. Baleanu, J. F. Alzaidy, E. Hashemizadeh, Operational matrix approach for solving the variable-order nonlinear Galilei invariant advection–diffusion equation, Advances in Difference Equations, 2018 (1), 1–11.
    [33] Y. Yang, Y. Ma, L. Wang, Legendre polynomials operational matrix method for solving fractional partial differential equations with variable coefficients, Math. Probl. Eng., 2015 (2015), 915195. https://doi.org/10.1155/2015/915195 doi: 10.1155/2015/915195
    [34] K. Shah, F. Jarad, T. Abdeljawad, Stable numerical results to a class of time-space fractional partial differential equations via spectral method, J. Adv. Res., 25 (2020), 39–48. https://doi.org/10.1016/j.jare.2020.05.022 doi: 10.1016/j.jare.2020.05.022
    [35] I. Singh, S. Arora, S. Kumar, Numerical solution of wave equation using Haar wavelet, Int. J.Pure and Appl. Math., 98 (2015), 457–469.
    [36] Y. Wang, Y. Chen, Shifted Legendre Polynomials algorithm used for the dynamic analysis of viscoelastic pipes conveying fluid with variable fractional order model, Appl. Math. Model., 81 (2020), 159–176. https://doi.org/10.1016/j.apm.2019.12.011 doi: 10.1016/j.apm.2019.12.011
  • This article has been cited by:

    1. Fang Wang, Lei Liu, Gennaro Infante, The Existence and Uniqueness of Solutions for Variable-Order Fractional Differential Equations with Antiperiodic Fractional Boundary Conditions, 2022, 2022, 2314-8888, 1, 10.1155/2022/7663192
    2. Khalid K. Ali, Nuha Al-Harbi, Abdel-Haleem Abdel-Aty, Traveling wave solutions to (3 + 1) conformal time derivative generalized q-deformed Sinh-Gordon equation, 2023, 65, 11100168, 233, 10.1016/j.aej.2022.10.020
    3. Atimad Harir, Said Melliani, Lalla Saadia Chadli, 2023, Chapter 7, 978-1-80356-566-8, 10.5772/intechopen.105904
    4. Ervin K. Lenzi, Luiz R. Evangelista, Haroldo V. Ribeiro, Richard L. Magin, Schrödinger Equation with Geometric Constraints and Position-Dependent Mass: Linked Fractional Calculus Models, 2022, 4, 2624-960X, 296, 10.3390/quantum4030021
    5. M. Adel, Dumitru Baleanu, Umme Sadiya, Mohammad Asif Arefin, M. Hafiz Uddin, Mahjoub A. Elamin, M.S. Osman, Inelastic soliton wave solutions with different geometrical structures to fractional order nonlinear evolution equations, 2022, 38, 22113797, 105661, 10.1016/j.rinp.2022.105661
    6. Hong Yan Xu, Hong Li, Xin Ding, Entire and meromorphic solutions for systems of the differential difference equations, 2022, 55, 2391-4661, 676, 10.1515/dema-2022-0161
    7. Mohammad Izadi, Şuayip Yüzbaşı, Waleed Adel, Accurate and efficient matrix techniques for solving the fractional Lotka–Volterra population model, 2022, 600, 03784371, 127558, 10.1016/j.physa.2022.127558
    8. Gurpreet Singh, Inderdeep Singh, Semi-analytical solutions of three-dimensional (3D) coupled Burgers’ equations by new Laplace variational iteration method, 2022, 6, 26668181, 100438, 10.1016/j.padiff.2022.100438
    9. Mamta Kapoor, Shehu transform on time-fractional Schrödinger equations – an analytical approach, 2022, 0, 1565-1339, 10.1515/ijnsns-2021-0423
    10. Valiyollah Ghazanfari, Mohammad Mahdi Shadman, Numerical solution of a comprehensive form of convection–diffusion equation for binary isotopes in the gas centrifuge, 2022, 175, 03064549, 109220, 10.1016/j.anucene.2022.109220
    11. Mamta Kapoor, Nehad Ali Shah, Wajaree Weera, Analytical solution of time-fractional Schrödinger equations via Shehu Adomian Decomposition Method, 2022, 7, 2473-6988, 19562, 10.3934/math.20221074
    12. Khalid Abdulaziz Alnowibet, Imran Khan, Karam M. Sallam, Ali Wagdy Mohamed, An efficient algorithm for data parallelism based on stochastic optimization, 2022, 61, 11100168, 12005, 10.1016/j.aej.2022.05.052
    13. Shao-Wen Yao, Sidheswar Behera, Mustafa Inc, Hadi Rezazadeh, Jasvinder Pal Singh Virdi, W. Mahmoud, Omar Abu Arqub, M.S. Osman, Analytical solutions of conformable Drinfel’d–Sokolov–Wilson and Boiti Leon Pempinelli equations via sine–cosine method, 2022, 42, 22113797, 105990, 10.1016/j.rinp.2022.105990
    14. Haifeng Wang, Yufeng Zhang, Application of Riemann–Hilbert method to an extended coupled nonlinear Schrödinger equations, 2023, 420, 03770427, 114812, 10.1016/j.cam.2022.114812
    15. Gacem Ilhem, Mahiéddine Kouche, Bedr'eddine Ainseba, Stability analysis of a fractional‐order SEIR epidemic model with general incidence rate and time delay, 2023, 0170-4214, 10.1002/mma.9161
    16. Asifa Tassaddiq, Sania Qureshi, Amanullah Soomro, Evren Hincal, Asif Ali Shaikh, A new continuous hybrid block method with one optimal intrastep point through interpolation and collocation, 2022, 2022, 2730-5422, 10.1186/s13663-022-00733-8
    17. A.A. Elsadany, Sarbast Hussein, A. Al-khedhairi, Amr Elsonbaty, On dynamics of 4-D blinking chaotic system and voice encryption application, 2023, 70, 11100168, 701, 10.1016/j.aej.2023.03.024
    18. Kezheng Zhang, Azzh Saad Alshehry, Noufe H. Aljahdaly, Rasool Shah, Nehad Ali Shah, Mohamed R. Ali, Efficient computational approaches for fractional-order Degasperis-Procesi and Camassa–Holm equations, 2023, 50, 22113797, 106549, 10.1016/j.rinp.2023.106549
    19. Samia Bushnaq, Asif Ullah Hayat, Hassan Khan, Numerical simulation of time-dependent viscous fluid flow with upward and downward fluctuation of spinning disk, 2024, 42, 2175-1188, 1, 10.5269/bspm.63089
    20. Ahmad El-Ajou, Rania Saadeh, Moawaih Akhu Dunia, Ahmad Qazza, Zeyad Al-Zhour, A new approach in handling one-dimensional time-fractional Schrödinger equations, 2024, 9, 2473-6988, 10536, 10.3934/math.2024515
    21. Farman Ali Shah, Zareen A Khan, Fatima Azmi, Nabil Mlaiki, A hybrid collocation method for the approximation of 2D time fractional diffusion-wave equation, 2024, 9, 2473-6988, 27122, 10.3934/math.20241319
    22. Nayereh Tanha, Behrouz Parsa Moghaddam, Mousa Ilie, Cutting-Edge Computational Approaches for Approximating Nonlocal Variable-Order Operators, 2024, 12, 2079-3197, 14, 10.3390/computation12010014
    23. Boddu Muralee Bala Krushna, Mahammad Khuddush, On the solvability of boundary value problems for iterative fractional differential equations, 2024, 73, 0009-725X, 1139, 10.1007/s12215-023-00975-4
    24. Allaoua Mehri, Hakima Bouhadjera, Mohammed S. Abdo, Hadeel Z. Alzumi, Wafa Shammakh, Finite element method for fractional order parabolic obstacle problem with nonlinear source term, 2024, 10, 26668181, 100721, 10.1016/j.padiff.2024.100721
    25. M.H. Heydari, A cardinal approach for two-dimensional modified anomalous space–time fractional sub-diffusion equation, 2023, 49, 22113797, 106545, 10.1016/j.rinp.2023.106545
    26. Sadam Hussain, Muhammad Sarwar, Nabil Mlaiki, Fatima Azmi, Existence and controllability of fractional semilinear mixed Volterra-Fredholm integro differential equation, 2023, 73, 11100168, 259, 10.1016/j.aej.2023.04.029
    27. S Naveen, V Parthiban, Variable‐order Caputo derivative of LC and RC circuits system with numerical analysis, 2024, 0098-9886, 10.1002/cta.4240
    28. Samia Bushnaq, Muhammad Sarwar, Hussam Alrabaiah, Existence theory and numerical simulations of variable order model of infectious disease, 2023, 19, 25900374, 100395, 10.1016/j.rinam.2023.100395
    29. Xing Fang, Leijie Qiao, Fengyang Zhang, Fuming Sun, Explore deep network for a class of fractional partial differential equations, 2023, 172, 09600779, 113528, 10.1016/j.chaos.2023.113528
    30. M. Abdelhakem, Dina Abdelhamied, M. El-kady, Y. H. Youssri, Enhanced shifted Tchebyshev operational matrix of derivatives: two spectral algorithms for solving even-order BVPs, 2023, 69, 1598-5865, 3893, 10.1007/s12190-023-01905-4
    31. Murugesan Manigandan, Saravanan Shanmugam, Mohamed Rhaima, Elango Sekar, Existence of Solutions for Caputo Sequential Fractional Differential Inclusions with Nonlocal Generalized Riemann–Liouville Boundary Conditions, 2024, 8, 2504-3110, 441, 10.3390/fractalfract8080441
    32. A N Nirmala, S Kumbinarasaiah, Nehad Ali Shah, Mansoor Alshehri, A new graph-theoretic approach for the study of the surface of a thin sheet of a viscous liquid model, 2023, 98, 0031-8949, 095248, 10.1088/1402-4896/acecb1
    33. Obaid J. Algahtani, Theoretical and numerical analysis of nonlinear Boussinesq equation under fractal fractional derivative, 2023, 12, 2192-8029, 10.1515/nleng-2022-0338
    34. Xiao Lan Liu, Hong Yan Xu, Yi Hui Xu, Nan Li, Results on solutions of several systems of the product type complex partial differential difference equations, 2024, 57, 2391-4661, 10.1515/dema-2023-0153
    35. Meriem Boukhobza, Amar Debbouche, Lingeshwaran Shangerganesh, Delfim F.M. Torres, Modeling the dynamics of the Hepatitis B virus via a variable-order discrete system, 2024, 184, 09600779, 114987, 10.1016/j.chaos.2024.114987
    36. Arzu Turan-Dincel, Sadiye Nergis Tural-Polat, Operational matrix method approach for fractional partial differential-equations, 2024, 99, 0031-8949, 125254, 10.1088/1402-4896/ad8f7a
    37. Mohamed S. Algolam, Osman Osman, Arshad Ali, Alaa Mustafa, Khaled Aldwoah, Amer Alsulami, Fixed Point and Stability Analysis of a Tripled System of Nonlinear Fractional Differential Equations with n-Nonlinear Terms, 2024, 8, 2504-3110, 697, 10.3390/fractalfract8120697
    38. Hicham Saber, Arshad Ali, Khaled Aldwoah, Tariq Alraqad, Abdelkader Moumen, Amer Alsulami, Nidal Eljaneid, Exploring Impulsive and Delay Differential Systems Using Piecewise Fractional Derivatives, 2025, 9, 2504-3110, 105, 10.3390/fractalfract9020105
  • Reader Comments
  • © 2022 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(2068) PDF downloads(87) Cited by(38)

Figures and Tables

Figures(6)  /  Tables(4)

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog