Abstract
We develop a local discontinuous Galerkin finite element method for the distributed-order time and Riesz space-fractional convection–diffusion and Schrödinger-type equations. The stability of the presented schemes is proved and optimal order of convergence \(\mathcal {O}(h^{N+1}+(\Delta t)^{1+\frac{\theta }{2}}+\theta ^{2})\) for the Riesz space-fractional diffusion and Schrödinger-type equations with distributed order in time, an order of convergence of \(\mathcal {O}(h^{N+\frac{1}{2}}+(\Delta t)^{1+\frac{\theta }{2}}\) \(+\theta ^{2})\) is provided for the Riesz space-fractional convection–diffusion equations with distributed order in time where h, \(\theta \) and \(\Delta t\) are space step size, the distributed-order variables and the step sizes in time, respectively. Finally, the performed numerical examples confirm the optimal convergence order and illustrate the effectiveness of the method.
Similar content being viewed by others
Explore related subjects
Discover the latest articles, news and stories from top researchers in related subjects.Avoid common mistakes on your manuscript.
1 Introduction
The distributed-order fractional differential equation (FDE), as a natural generalization of the single-order and the multi-term FDE, has become more popular in many physical and biological applications recently, for example, dielectric induction and diffusion [1], viscoelastic oscillators [2], the stress behavior of an elastic medium [3], to study properties of rheological of composite materials [4, 5], distributed-order membranes in the ear [6]. This kind of equations was first introduced by Caputo [9] who used to find infinite plates and dielectric spherical shells or the eigenfunctions of the torsional models of anelastic. FDEs have recently been shown as a good tool to model the complex dynamical systems than the classical and fractional-order models. Moreover, the time-fractional anomalous diffusion equation with a constant order temporal derivative cannot be used to model the processes lacking temporal scaling, so in some sense, the diffusion equation with distributed order in time can be used to describe processes getting more anomalous in course of time, e.g., the accelerating superdiffusion and retarded sub-diffusion [7, 8].
In recent years, there has been much attention and effort put on developing various numerical algorithms for solving distributed-order and space-fractional equations. Diethelm and Ford [10, 11] transformed the distributed-order ordinary differential equations into a multi-term FDEs by quadrature formula. Ford and Morgado [12] show the uniqueness, existence and the approximation of the solution of distributed FDEs. Podlubny et al. [13] solved the distributed-order FDE by a matrix approach method. Ye et al. [14] discussed the Riesz space-fractional diffusions equations with distributed order in time by an implicit difference scheme. They prove that the implicit difference scheme is convergent and unconditionally stable. In [15], the authors have developed Jacobi collocation method for the multi-dimensional distributed-order generalized Schrödinger equations [15]. There exist many numerical methods for solving partial differential equations (PDEs) and fractional partial differential equations (FPDEs) in [16,17,18,19,20,21,22] and references therein. They differ in accuracy, performance and applicability. To the best of our knowledge, however, the LDG method, which is an important approach to solve FPDEs, has not been considered for the time distributed-order FPDEs. Here, we develop a LDG method to solve the time distributed-order and Riesz space-fractional convection–diffusion equation (TDO-RSFCDE)
the nonlinear time distributed-order and Riesz space-fractional Schrödinger equation (TDO-RSFSE)
and the coupled nonlinear TDO-RSFSEs
where homogeneous boundary conditions and \(x\in \mathbb {R}\), \(t\in (0,T]\). \(\varepsilon , \varepsilon _{i}\), \(i=1,2,3,4\) are a real constants and g(u) and f(u) are nonlinear real functions, and \(\mathcal {D}_{t}^{W(\alpha )}{}u(x,t)\) denotes the distributed-order fractional derivative of u in time t, given by
where \(W(\alpha )\) is the weight function, \({}^{\,\,\,C}_{\,\,\,\,0}\mathcal {D}_{t}^{\alpha }u(x,t)\) \((0< \alpha < 1)\) is the Caputo fractional derivative of order \(\alpha \) with respect to t. The fractional Laplacian \(-(-\Delta )^{\frac{\beta }{2}}\) can be characterized using Fourier analysis as [23,24,25]
where \(\mathcal {F}\) is the Fourier transform.
The Discontinuous Galerkin (DG) method is famous for high accuracy properties and extreme flexibility [26,27,28,29]. There exist many applications of DG methods to solve FPDEs, for example, time-fractional wave and diffusion equations [30,31,32,33], fractional convection–diffusion equations [34, 35], fractional Cahn–Hilliard equation [36] and the nonlinear Riesz space-fractional Schrödinger-type equations [35, 37].
The organization of the paper is as follows: we present some preliminary definitions and a few lemmas in Sect. 2. In section 3, we present the DG scheme for the TDO-RSFCDEs. Then we prove \(L^{2}\) stability as well as an error estimate in Sect. 4. In Sect. 5, we present a LDG method for the nonlinear distributed-order time and Riesz space-fractional Schrödinger-type equations. We introduce the LDG scheme for the TDO-RSFSE in Sect. 5.1. Moreover, we give \(L^{2}\) stability for the nonlinear case in Sect. 5.1.1 and an error estimations for the linear equation in Sect. 5.1.2. In Sect. 5.2, we introduce the LDG scheme for the coupled nonlinear TDO-RSFSEs and give a theoretical result of \(L^{2}\) stability and error estimates. Section 6 presents some numerical experiments to illustrate the efficiency of the scheme. Finally, some conclusions are given in Sect. 7.
2 Preliminaries
We first introduce the definitions of fractional derivatives and integrals [38] and review a few lemmas for our analysis.
2.1 Caputo–Liouville fractional calculus
The right-sided and left-sided Riemann–Liouville (R–L) integrals of order \(\mu \), when \(0< \mu < 1\), for the function f(x) is defined , respectively, as
and
For any \(n\in \mathbb {N^{+}} ,(n-1<\mu <n)\), the right and left R–L fractional derivatives of function f are defined by
and
Furthermore, the lift-sided and right-sided Caputo derivatives of function f are obtained as
and
Proposition 1
The R–L fractional derivatives of a function f relates to the Caputo definitions by the following equation
which implies \(\big ({}^{\,\,\,\,\,\,C}_{-a}\mathcal {D}_{x}^{\mu }f\big )(x) =\big ({}^{\,\,RL}_{-a}\mathcal {D}_{x}^{\mu }f\big )(x)\), if \(f^{(j)}(a)=0\), \(j=0,\ldots ,n-1\).
Proposition 2
Semigroup and linearity property
The Riesz fractional derivative is given by
If \(\mu <0\), the fractional Laplacian becomes the fractional integral operator. In this case, for any \(0<\alpha <1\), which is formally defined as follows
When \(1<\mu <2\), using (2.5), (2.6) and (2.11), we can rewrite the fractional Laplacian in the following form:
Definition 1
(The right, left and symmetric fractional spaces [39]) For any \(0<\alpha <1\), we define the seminorms
and the norms
and let the three spaces \(J_{R}^{\alpha }(\mathbb {R})\), \(J_{L}^{\alpha }(\mathbb {R})\) and \(J_{S}^{\alpha }(\mathbb {R})\) denote the closures of \(C_{0}^{\infty }(\mathbb {R})\) with respect to \(\Vert .\Vert _{J_{R}^{\alpha }(\mathbb {R})}\) and \(\Vert .\Vert _{J_{L}^{\alpha }(\mathbb {R})}\), respectively.
Lemma 1
(see [39]) For \(\alpha >0\), assume that f(x) is a real function. Then
Generally, we assume that the fractional problems in a bounded domain and let the domain \(\varOmega = [a, b]\) instead of \(\mathbb {R}\).
Lemma 2
(fractional Poincaré–Friedrichs, [39]) If \(\alpha \in \mathbb {R}\), we have
and
Lemma 3
(See [40]) The fractional integration operator \(\mathcal {I}^{s}\) is bounded in \(L^{2}(\varOmega )\):
where \(\mathcal {I}^{\alpha }={}^{RL}_{\,x_{L}}\mathcal {I}_{x}^{\alpha }\) (i.e., right-sided R–L integral of order \(\alpha \)).
Lemma 4
(See [37]) The fractional integration operator \(\Delta _{-\alpha }\) is bounded in \(L^{2}(\varOmega )\):
3 LDG scheme for the TDO-RSFCDE
Let us consider the TDO-RSFCDE. To discretize the integral interval [0, 1], we choose the grid \(0=\tau _{0}<\tau _{1}<\cdots <\tau _{S}=1\) and take \(\Delta \tau _{j}=\tau _{j}-\tau _{j-1}=\frac{1}{S}=\theta \), \(\alpha _{j}=\frac{\tau _{j}+\tau _{j-1}}{2}=\frac{2j-1}{2S}\), \(j=1,2,\ldots ,S\), \(S\in \mathbb {N}\) and by the mid-point quadrature rule, we get
Thus the TDO-RSFCDE (1.1) is expressed as multi-term fractional equation. We approximate the time-fractional derivative (3.1) by simple quadrature formula in [41]. Let \(\Delta t=T/M\) be the time mesh size, \(t_{n}=n\Delta t,n=0,1,\ldots ,M\) be mesh points and M is a positive integer.
Lemma 5
(See [41]) Let \(z(t)\in \mathcal {C}^{2}[0,t_{n}]\), for any \(0<\alpha <1\). It holds that
We suppose the notation
From (1.4), (3.1) and (3.3), we obtain
where \(\lambda _{j}=(\Delta t)^{\alpha _{j}} \varGamma (2-\alpha _{j})\) and \(a_{l}^{\alpha _{j}}=(l+1)^{1-\alpha _{j}}-l^{1-\alpha _{j}}\).
For space-fractional derivative, we rewrite it as a composite of a fractional integral and first-order derivatives and convert the TDO-RSFCDE (1.1) into a system of low order equations. However, for the first-order system, alternating fluxes are used. We suppose three variables r, p and q and set
then, the TDO-RSFCDE can be rewritten as
For any real number r, the broken Sobolev space is defined as
The \(L^{2}(D^{k})\) norm and local inner product are defined by
and the norm and global broken inner product
We introduce some notation
For simplicity, we discretize the computational domain \(\varOmega \) into K non-overlapping elements, \(D^{k}=[x_{k-\frac{1}{2}},x_{k+\frac{1}{2}}]\), \(k = 1,\ldots ,K\). Let \(u_{h}^{n}, p_{h}^{n}, q_{h}^{n}, r_{h}^{n}\in V_{k}^{N}\) be the approximation of \(u(.,t_{n}), p(.,t_{n}),q(.,t_{n}),r(.,t_{n})\), respectively, where the approximation space is defined as
where \(\mathbb {P}(D^{k})\) indicate the set of degree polynomials up to N defined on the element \(D^{k}\).
We define a fully discrete LDG scheme with as follows: find \(u_{h}^{n},p_{h}^{n}, q_{h}^{n}, r_{h}^{n}\in V_{k}^{N}\), such that for all test functions \(v,\psi ,\phi ,\eta \in V_{k}^{N}\),
We can choose numerical fluxes as
Note that we can also choose
4 Stability analysis and error estimates
Here, we focus on providing the stability analysis and the error estimates of the proposed scheme, for TDO-RSFCDE.
4.1 The analysis of stability for fully discrete scheme
Theorem 1
The fully discrete LDG scheme (3.12) is stable, and
Proof
See “Appendix A”. \(\square \)
4.2 Error estimates
we discuss the optimal error for the LDG scheme (3.12). We want to first get the equation of error.
The exact solution of (1.1) satisfies
Subtracting equation (3.12) from (4.2), we can obtain the error equation
where
such that
For the error estimate, we define special projections, \(\mathcal {P}\) and \(\mathcal {P}^{\pm }\) into \(V_{h}^{k}\). For all the elements, \(D^{k}\), \(k = 1, 2, \ldots , K\) are defined to satisfy
Denoting
From the standard approximation theory [42], the projections satisfy the following inequality
where \(\pi ^{e}=\mathcal {P}^{\pm }u^{n}-u_{h}^{n}\) or \(\pi ^{e}=\mathcal {P}u^{n}-u_{h}^{n}\). \(\varGamma _{h}\) denotes the set of boundary points of all elements \(D^{k}\). The positive constant C, solely depending on \(u^{n}\), is independent of h.
Theorem 2
(Diffusion without convection \(f(u)=0\)) The \(L^{2}\) error of the scheme (3.12) satisfies:
Proof
See “Appendix B”.
For the TDO-RSFCDE, we review a few lemmas for our analysis. \(\square \)
Lemma 6
(see [43]) For any piecewise smooth function \(\pi \in L^{2}(\varOmega )\), on each cell boundary point we define
The nonlinear part can be rewritten as
We can rewrite (4.11) as:
Lemma 7
(see [43]) For \(\mathcal {H}_{k}(f;u,u_{h};\pi )\) defined above, we have the following estimate:
and
Theorem 3
The \(L^{2}\) error of the scheme (3.12) satisfies:
Proof
See “Appendix C”. \(\square \)
5 LDG method for the nonlinear distributed-order time and Riesz space-fractional Schrödinger-type equations
5.1 LDG method for the nonlinear TDO-RSFSE
We introduce three variables e, r, s and set
then, the nonlinear TDO-RSFSE can be rewritten as
We set the complex function u(x, t) as
so, the nonlinear TDO-RSFSE (5.2) can be rewritten as
Let \(p_{h}^{n}, q_{h}^{n}, e_{h}^{n},l_{h}^{n},r_{h}^{n},s_{h}^{n},w_{h}^{n},z_{h}^{n}\in V_{k}^{N}\) be the approximation of \(p(.,t_{n}),q(.,t_{n}),r(.,t_{n}),s(.,t_{n}),e(.,t_{n}),l(.,t_{n}),\) \( w(.,t_{n}),z(.,t_{n})\), respectively.
Let \(p_{h}^{n}, q_{h}^{n},r_{h}^{n},s_{h}^{n},e_{h}^{n},l_{h}^{n},w_{h}^{n},z_{h}^{n}\in V_{k}^{N}\), and test functions \(\vartheta _{1},\rho ,\phi ,\varphi ,\chi ,\varrho ,\psi ,\zeta \in V_{k}^{N}\),
The numerical traces \(( \widehat{p}^{n}_{h}, \widehat{q}^{n}_{h},\widehat{s}^{n}_{h},\widehat{z}^{n}_{h})\) are defined on interelement faces as the alternating fluxes [22, 44]:
Note that we can also choose
5.1.1 The analysis of stability for fully discrete scheme
we focus on providing the stability analysis of the proposed scheme, for TDO-RSFSE
Theorem 4
(\(L^{2}\) stability) The semidiscrete scheme (5.5) is stable, and
Proof
We can prove by similar techniques as that in the proof of Theorem 4.1. \(\square \)
5.1.2 Error estimates
Here, we focus on providing the error estimates of the proposed scheme, for TDO-RSFSE. We consider the linear TDO-RSFSE
The exact solution of (5.9) satisfies
Subtracting (5.5) from (5.10), we can obtain the error equation
Denoting
Lemma 8
where
Proof
See “Appendix D”. \(\square \)
Theorem 5
The \(L^{2}\) error of the scheme (5.5) satisfies:
Proof
We estimate the term \(Q_{i}\), \(i=1,\ldots ,4\) in Lemma 8. So we employ Young’s inequality, Lemma 4 and the approximation results (4.8) and similar techniques as that in the proof of Theorem 2, we obtain the result. \(\square \)
5.2 LDG method for the coupled nonlinear TDO-RSFSEs
We present the LDG method for the coupled nonlinear TDO-RSFSEs
We rewrite (5.16) as a first-order system:
Setting \(u_{1}(x, t)=p(x, t) + iq(x, t)\) and \(u_{2}(x, t) = \upsilon (x, t) + i\vartheta (x, t)\) in system (5.16), we obtain the following coupled system
Let \(p_{h}^{n}, q_{h}^{n}, Q^{n},r_{h}^{n},s_{h}^{n},H^{n}_{h},w_{h}^{n},z_{h}^{n}\), \(\upsilon _{h}^{n},\vartheta _{h}^{n}, L^{n}_{h},\rho _{h}^{n},\varpi _{h}^{n},E_{h}^{n},\xi _{h}^{n}\), \(\varrho _{h}^{n} \in V_{k}^{N}\), and test functions \(\vartheta _{1},\beta _{1},\phi ,\varphi ,\chi ,\beta _{2},\psi \), \(\zeta ,\gamma ,\beta _{3},\delta ,\varsigma ,o,\beta _{4},\omega ,\kappa \in V_{k}^{N}\),
The numerical traces \(( \widehat{p}^{n}_{h}, \widehat{q}^{n}_{h}, \widehat{s}^{n}_{h},\widehat{z}^{n}_{h},\widehat{\upsilon }^{n}_{h}, \widehat{\vartheta }^{n}_{h},\widehat{\varpi }^{n}_{h},\widehat{\varrho } ^{n}_{h})\) are defined as
Theorem 6
(\(L^{2}\) stability) Let \(u_{1h}^{n}, u_{2h}^{n}\in V_{k}^{N}\) be the approximation of \(u_{1}(x,t_{n}),u_{2}(x,t_{n})\), then the solution to the scheme (5.19) and (5.20) satisfies the \(L^{2}\) stability
Theorem 7
Let \(u_{1}(x,t_{n})\) and \(u_{2}(x,t_{n})\) be the exact solutions of the linear coupled TDO-RSFSEs (5.16), and let \(u_{1h}^{n}\) and \(u_{2h}^{n}\) be the numerical solutions of the scheme (5.19). Then the \(L^{2}\) error satisfies:
Theorems 7 and 6 can be proven by similar techniques as that in the proof of Theorems 4 and 5. We will thus not give the details here.
6 Numerical examples
In the following, we present some numerical experiments to show the accuracy and the performance of the present LDG method for the space-fractional convection–diffusion and Schrödinger-type equations with distributed order in time.
Example 1
Consider the following Riesz space-fractional diffusion equation with distributed-order in time
and the forcing term
then the exact solution is given by \(u(x,t)=t^{2}(x^{2}-1)^{4}\) with \(\varepsilon =\frac{\varGamma (8-\beta )}{\varGamma (8)}\).
The spatial convergence orders and \(L^{2}\) errors are listed in Table 1. The numerical of spatial convergence orders are shown in Fig. 1 with \(N = 5,6, 7\) and \(K = 5, 10, 15, 20, 25\). From these results, we show that the spatial convergence order is \(\mathcal {O}(h^{N+1})\). Figure 2 exhibits a comparison between the exact and numerical solution that show the efficiency of the simulation.
Example 2
We consider the following Riesz space-fractional Burgers’ equation with distributed-order in time
and the forcing term
and the exact solution will be \(u(x,t)=t^{2}(x^{2}-1)^4\) with \(\varepsilon =\frac{\varGamma (8-\beta )}{\varGamma (8)}\). For the nonlinear term, we select a Lax–Friedrichs flux and take \(\Delta t=T/500,\,\theta =1/50\). The spatial convergence orders and \(L^{2}\) errors are listed in Table 2. Table 3 provides the temporal convergence orders and the \(L^{2}\) errors with \(\beta = 1.2, 1.6\), respectively, at \(T = 0.5\) with \(N=1\), \(K=30\). The numerical integration convergence orders and \(L^{2}\) errors in Table 4 with \(\beta = 1.2, 1.6\), respectively, at \(T = 0.5\). From these results, we show that the order of convergence is \(\mathcal {O}(h^{N+1}+(\Delta t)^{1+\frac{\theta }{2}}+\theta ^{2})\) when \(\theta \) is small enough.
Example 3
Consider the following nonlinear TDO-RSFSE
and the forcing term
The exact solution \(u(x,t)=(1+i)t^{2}(x^{2}-1)^5\) with \(\varepsilon =\frac{\varGamma (10-\beta )}{\varGamma (10)}\). The spatial convergence orders and \(L^{2}\) errors are listed in Table 5. The numerical spatial convergence orders are shown in Fig. 3 with \(N = 3,4, 5\) and \(K = 15, 20, 25, 30\). Table 6 provides the temporal convergence orders and the \(L^{2}\) errors with \(\beta = 1.2, 1.6\), respectively, at \(T = 0.5\). The numerical integration convergence orders and \(L^{2}\) errors in Table 7 with \(\beta = 1.2, 1.6\), respectively, at \(T = 0.5\). From these results it is clear that the order of convergence is \(\mathcal {O}(h^{N+1}+(\Delta t)^{1+\frac{\theta }{2}}+\theta ^{2})\), which matches the theoretical convergence order when \(\theta \) is small enough.
Example 4
We consider the coupled nonlinear TDO-RSFSEs
and the forcing terms
to obtain an exact solutions \(u_{1}(x,t)=(1+i)t^{2}(x^{2}-1)^6\) and \(u_{2}(x,t)=(1+i)t^{2}(x^{2}-1)^6\) with \(\beta =1.3,\,\varepsilon _{1}=\frac{\varGamma (13-\beta )}{2\varGamma (13)}\), \(\varepsilon _{2}=\frac{\varGamma (13-\beta )}{2\varGamma (13)}\). The spatial convergence orders and \(L^{2}\) errors are listed in Tables 8 and 9, confirming optimal \(\mathcal {O}(h^{N+1})\) order of convergence across.
7 Conclusions
In this work, we developed and analyzed a LDG method for solving the distributed-order time and Riesz space-fractional convection–diffusion and Schrödinger-type equations and have proven the stability and error estimates of these methods. The performed numerical experiments confirm the optimal convergence order. Future work will include the analysis of LDG method for two-dimensional fractional problems.
References
Caputo, M.: Distributed order differential equations modelling dielectric induction and diffusion. Fract. Calc. Appl. Anal. 4, 421–442 (2001)
Atanackovic, T., Budincevic, M., Pilipovic, S.: On a fractional distributed-order oscillator. J. Phys. A Math. Gen. 38, 6703 (2005)
Caputo, M.: Linear models of dissipation whose Q is almost frequency independentII. Geophys. J. Int. 13, 529–539 (1967)
Hartley, T.T., Lorenzo, C.F.: Fractional system identification: an approach using continuous order-distributions. NASA/TM-1999-209640 (1999)
Lorenzo, C.F., Hartley, T.T.: Variable order and distributed order fractional operators. Nonlinear Dyn. 29, 57–98 (2002)
Naghibolhosseini, M.: Estimation of outer-middle ear transmission using DPOAEs and fractional-order modeling of human middle ear Ph.D thesis, City University of New York, NY (2015)
Caputo, M.: Diffusion with space memory modelled with distributed order space fractional differential equations. Ann. Geophys. 46, 223–234 (2003)
Sokolov, A.C.I., Klafter, J.: Distributed order fractional kinetics. Acta Phys. Pol. B 35, 1323–1341 (2004)
Caputo, M.: Elasticitàe Dissipazione. Zanichelli, Bologna (1969)
Diethelm, K., Ford, N.J.: Numerical analysis for distributed-order differential equations. J. Comput. Appl. Math. 225, 96–104 (2009)
Diethelm, K., Ford, N.J.: Numerical solution methods for distributed order differential equations. Fract. Calc. Appl. Anal. 4, 531–542 (2001)
Ford, N.J., Morgado, M.L.: Distributed order equations as boundary value problems. Comput. Math. Appl. 64, 2973–2981 (2012)
Podlubny, I., Skovranek, T., Jara, B.M.V., Petras, I., Verbitsky, V., Chen, Y.: Matrix approach to discrete fractional calculus III: non-equidistant grids, variable step length and distributed orders. Philos. Trans. R. Soc. A 371, 20120153 (2013)
Ye, H., Liu, F., Anh, V., Turner, I.: Numerical analysis for the time distributed-order and Riesz space fractional diffusions on bounded domains. IMA J. Appl. Math. 80, 825–838 (2013)
Bhrawy, A., Zaky, M.: Numerical simulation of multi-dimensional distributed-order generalized Schrödinger equations. Nonlinear Dyn. 89, 1415–1432 (2017)
Shen, J., Tang, T., Wang, L.-L.: Spectral Methods: Algorithms, Analysis and Applications, vol. 41. Springer, New York (2011)
Zeng, S., Migórski, S.: A class of time-fractional hemivariational inequalities with application to frictional contact problem. Commun. Nonlinear Sci. Numer. Simul. 56, 34–48 (2018)
Mahmoud, G.M., Ahmed, M.E., Abed-Elhameed, T.M.: On fractional-order hyperchaotic complex systems and their generalized function projective combination synchronization. Optik Int. J. Light Electron Opt. 130, 398–406 (2017)
Yüzbaşı, Ş., Karaçayır, M.: An exponential galerkin method for solutions of hiv infection model of \({CD4}^{+}\) T-cells. Comput. Biol. Chem. 67, 205–212 (2017)
Mahmoud, G.M., Abed-Elhameed, T.M., Ahmed, M.E.: Generalization of combination-combination synchronization of chaotic n-dimensional fractional-order dynamical systems. Nonlinear Dyn. 83, 1885–1893 (2016)
Yüzbaşı, Ş., Karaçayır, M.: A galerkin-like approach to solve high-order integrodifferential equations with weakly singular kernel. Kuwait J. Sci. 43, 106–120 (2016)
Cockburn, B., Shu, C.-W.: The local discontinuous Galerkin method for time-dependent convection-diffusion systems. SIAM J. Numer. Anal. 35, 2440–2463 (1998)
El-Sayed, A., Gaber, M.: On the finite caputo and finite Riesz derivatives. Electron. J. Theor. Phys. 3, 81–95 (2006)
Muslih, S.I., Agrawal, O.P.: Riesz fractional derivatives and fractional dimensional space. Int. J. Theor. Phys. 49, 270–275 (2010)
Yang, Q., Liu, F., Turner, I.: Numerical methods for fractional partial differential equations with Riesz space fractional derivatives. Appl. Math. Model. 34, 200–218 (2010)
Bassi, F., Rebay, S.: A high-order accurate discontinuous finite element method for the numerical solution of the compressible Navier–Stokes equations. J. Comput. Phys. 131, 267–279 (1997)
Cockburn, B., Shu, C.-W.: TVB Runge–Kutta local projection discontinuous Galerkin finite element method for conservation laws. II. General framework. Math. Comput. 52, 411–435 (1989)
Cockburn, B., Dawson, C.: Approximation of the velocity by coupling discontinuous Galerkin and mixed finite element methods for flow problems. Comput. Geosci. 6, 505–522 (2002)
Cockburn, B., Kanschat, G., Schötzau, D.: A locally conservative LDG method for the incompressible Navier–Stokes equations. Math. Comput. 74, 1067–1095 (2005)
Mustapha, K., McLean, W.: Piecewise-linear, discontinuous Galerkin method for a fractional diffusion equation. Numer. Algorithms 56, 159–184 (2011)
Mustapha, K., McLean, W.: Uniform convergence for a discontinuous Galerkin, time-stepping method applied to a fractional diffusion equation. IMA J. Numer. Anal. 32, 906–925 (2012)
Mustapha, K., McLean, W.: Superconvergence of a discontinuous Galerkin method for fractional diffusion and wave equations. SIAM J. Numer. Anal. 51, 491–515 (2013)
Deng, W., Hesthaven, J.S.: Local discontinuous Galerkin methods for fractional diffusion equations. ESAIM Math. Model. Numer. Anal. 47, 1845–1864 (2013)
Xu, Q., Hesthaven, J.S.: Discontinuous Galerkin method for fractional convection-diffusion equations. SIAM J. Numer. Anal. 52, 405–423 (2014)
Aboelenen, T.: A direct discontinuous Galerkin method for fractional convection-diffusion and Schrödinger type equations. arXiv:1708.04546 (2017)
Aboelenen, T., El-Hawary, H.: A high-order nodal discontinuous Galerkin method for a linearized fractional Cahn–Hilliard equation. Comput. Math. Appl. 73, 1197–1217 (2017)
Aboelenen, T.: A high-order nodal discontinuous Galerkin method for nonlinear fractional Schrödinger type equations. Commun. Nonlinear Sci. Numer. Simul. 54, 428–452 (2018)
Miller, K., Ross, B.: An Introduction to the Fractional Calculus and Fractional Differential Equations. Wiley (1993). https://books.google.co.in/books?id=MOp_QgAACAAJ
Ervin, V.J., Roop, J.P.: Variational formulation for the stationary fractional advection dispersion equation. Numer. Methods Partial Differ. Equ. 22, 558–576 (2006)
Kilbas, A.A., Srivastava, H.M., Trujillo, J.J.: Theory and Applications of Fractional Differential Equations (North-Holland Mathematics Studies), vol. 204. Elsevier, New York (2006)
Sun, Z.-Z., Wu, X.: A fully discrete difference scheme for a diffusion-wave system. Appl. Numer. Math. 56, 193–209 (2006)
Ciarlet, P.G.: Finite Element Method for Elliptic Problems. Society for Industrial and Applied Mathematics, Philadelphia (2002)
Zhang, Q., Shu, C.-W.: Error estimates to smooth solutions of Runge–Kutta discontinuous Galerkin methods for scalar conservation laws. SIAM J. Numer. Anal. 42, 641–666 (2004)
Yan, J., Shu, C.-W.: Local discontinuous Galerkin methods for partial differential equations with higher order derivatives. J. Sci. Comput. 17, 27–47 (2002)
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix A: Proof Of Theorem 4.1
Set \((v,\psi ,\phi ,\eta )=(u_{h}^{n},p_{h}^{n}-q_{h}^{n}, u_{h}^{n},r_{h}^{n})\) in (3.12), and define \(\theta (u_{h}^{n})=\int ^{u_{h}^{n}}f(s_{h}^{n})\mathrm{d}s_{h}^{n}\). Then the following result holds:
Summing over k, with the numerical fluxes (3.13), we obtain
Here \(\widehat{f}((u_{h}^{n})^{-},(u_{h}^{n})^{+})\) is monotone flux, we have
This implies that
Employing Young’s inequality and Lemma 4, we obtain
Recalling Lemma 4, we obtain
It then follows that
Using Cauchy–Schwarz inequality, we obtain
where \(Q=\frac{1}{\sum _{j=1}^{S}\frac{W(\alpha _{j})\Delta \tau _{j}}{\lambda _{j}}}\) and provided c is sufficiently small such that \(1-cQ>0\), we obtain that
Clearly the theorem is true for \(n = 1\). It is also true for \(n = 1, 2, \ldots , m-1\). So, by (A.9), we have
\(\square \)
Appendix B: Proof Of Theorem 2
From (4.3), we can get the error equation
Using (4.7), the error equation (B.1) can be written
and taking the test functions
we obtain
by the projection properties \(P^{+}\) and \(P^{-}\) we obtain
Employing Young’s inequality and Lemma 4 and the property of interpolation (4.8) and (4.4), we obtain
by using Lemma 5, (4.7) and the interpolation property (4.8), we obtain
From (3.1), (4.5) and (B.7), we obtain
Hence
Recalling Lemma 2 and provided \(c_{i},\,\,i=1,2\) are sufficiently small such that \(c_{i}\le 1\), we obtain
It then follows that
Employing Young’s inequality, we obtain
Notice the facts that
Thus,
provided c is sufficiently small such that \(\frac{3}{4}-cQ>0\), we obtain that
Clearly the theorem is true for \(n = 0\). It is also true for \(n = 1, 2, \ldots , m-1\). Then, by (B.15), we have
Finally, by using triangle inequality and standard approximation theory, we get
\(\square \)
Appendix C: Proof Of Theorem 3
Using (4.7), the error equation (4.3) can be written
Following the proof of Theorem 2, we take the test functions
we obtain
by the projection properties \(P^{+}\) and \(P^{-}\), we obtain
Employing Young’s inequality and Lemma 4 and the property of interpolation (4.8), (B.7) and (B.8), we obtain
Recalling Lemma 2 and provided \(c_{i},\,\,i=1,2\) are sufficiently small such that \(c_{i}\le 1\), we obtain
It then follows that
Employing Young’s inequality, we obtain
provided c is sufficiently small such that \(\frac{3}{4}-cQ>0\), we obtain that
Clearly the theorem is true for \(n = 0\). It is also valid for \(n = 1, 2, \ldots , m-1\). Then, by (C.9), we have
Finally, by using triangle inequality and standard approximation theory, we can get (4.15). \(\square \)
Appendix D: Proof Of Lemma 8
From the Galerkin orthogonality (5.11), we get
We take the test functions
we obtain
Summing over k, simplify by integration by parts and (5.6). \(\square \)
Rights and permissions
About this article
Cite this article
Aboelenen, T. Local discontinuous Galerkin method for distributed-order time and space-fractional convection–diffusion and Schrödinger-type equations. Nonlinear Dyn 92, 395–413 (2018). https://doi.org/10.1007/s11071-018-4063-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11071-018-4063-y