Next Article in Journal
Processing the Controllability of Control Systems with Distinct Fractional Derivatives via Kalman Filter and Gramian Matrix
Next Article in Special Issue
Fractional Second-Grade Fluid Flow over a Semi-Infinite Plate by Constructing the Absorbing Boundary Condition
Previous Article in Journal
Deep-Learning Estimators for the Hurst Exponent of Two-Dimensional Fractional Brownian Motion
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

A Mixed Finite Element Method for the Multi-Term Time-Fractional Reaction–Diffusion Equations

1
School of Statistics and Mathematics, Inner Mongolia University of Finance and Economics, Hohhot 010070, China
2
School of Mathematical Sciences, Inner Mongolia University, Hohhot 010021, China
*
Author to whom correspondence should be addressed.
Fractal Fract. 2024, 8(1), 51; https://doi.org/10.3390/fractalfract8010051
Submission received: 5 December 2023 / Revised: 4 January 2024 / Accepted: 8 January 2024 / Published: 12 January 2024

Abstract

:
In this work, a fully discrete mixed finite element (MFE) scheme is designed to solve the multi-term time-fractional reaction–diffusion equations with variable coefficients by using the well-known L 1 formula and the Raviart–Thomas MFE space. The existence and uniqueness of the discrete solution is proved by using the matrix theory, and the unconditional stability is also discussed in detail. By introducing the mixed elliptic projection, the error estimates for the unknown variable u in the discrete L ( L 2 ( Ω ) ) norm and for the auxiliary variable λ in the discrete L ( ( L 2 ( Ω ) ) 2 ) and L ( H ( div , Ω ) ) norms are obtained. Finally, three numerical examples are given to demonstrate the theoretical results.

1. Introduction

Fractional calculus and fractional partial differential equations (FPDEs) have been confirmed to be very important tools in describing some anomalous phenomena and processes with memory and nonlocal properties [1,2,3,4,5,6]. Moreover, some underlying and complex processes can be described more appropriately by multi-term FPDEs [7,8,9], as they contains multiple fractional derivative or calculus terms. In recent years, many numerical methods have been increasingly used by scholars to solve multi-term FPDEs. Liu et al. [10] constructed some finite difference (FD) schemes to solve the multi-term time-fractional wave-diffusion equations by using two fractional predictor–corrector methods. Dehghan et al. [11] devised two high-order numerical schemes to solve the multi-term time-fractional diffusion-wave equations by using the compact FD method and Galerkin spectral technique. Ren and Sun [12] established an efficient compact FD scheme for the multi-term time-fractional diffusion-wave equation by using the L 1 formula. Zheng et al. [13] proposed a high-order space–time spectral method for the multi-term time-fractional diffusion equations by using the Legendre polynomials in the temporal direction and the Fourier-like basis functions in the spatial direction. Du and Sun [14] constructed an FD scheme for multi-term time-fractional mixed diffusion and wave equations by using the L 2 1 σ formula. Hendy and Zaky [15] proposed a spectral method for a coupled system of nonlinear multi-term time–space fractional diffusion equations by using the L 1 formula on a time-graded mesh. Liu et al. [16] developed an ADI Legendre spectral method for solving a multi-term time-fractional Oldroyd-B fluid-type diffusion equation. Wei and Wang [17] constructed a higher-order numerical scheme for the multi-term variable-order time-fractional diffusion equation by using the local discontinuous Galerkin method. She et al. [18] considered a spectral method for solving the multi-term time-fractional diffusion problem by using a modified L 1 formula.
Meanwhile, many scholars selected the finite element (FE) method for solving the multi-term FPDEs and have achieved excellent results. Jin et al. [19] developed an FE method for a multi-term time-fractional diffusion equation and considered the case of smooth and nonsmooth initial data. Li et al. [20] proposed an FE method to solve a higher-dimensional multi-term fractional diffusion equation on nonuniform time meshes. Zhou et al. [21] developed a weak Galerkin FE method for solving multi-term time-fractional diffusion equations by using a convolution quadrature formula. Bu et al. [22] proposed a space–time FE method for solving the multi-term time–space fractional diffusion equation based on the suitable graded time mesh. Feng et al. [23] proposed an FE method for a multi-term time-fractional mixed subdiffusion and diffusion-wave equation on the convex domain by using mixed L-type schemes. Meng and Stynes [24] considered an L 1 FE method for a multi-term time-fractional initial-boundary value problem on the temporal graded mesh. Yin et al. [25] constructed a class of efficient time-stepping FE schemes for multi-term time-fractional reaction–diffusion-wave equations by using the shifted convolution quadrature method. Huang et al. [26] proposed an α -robust FE method for a multi-term time-fractional diffusion problem on a graded mesh by using the L 1 formula. Liu et al. [27] proposed an FE method for solving a multi-term variable-order time-fractional diffusion equation and developed an efficient parallel-in-time algorithm to reduce the computational costs.
In this work, we will construct a fully discrete mixed finite element (MFE) scheme for the following multi-term time-fractional reaction–diffusion (TFRD) equations with variable coefficients:
P ( D t ) u ( x , t ) div ( A ( x ) u ( x , t ) ) + p ( x ) u ( x , t ) = f ( x , t ) , ( x , t ) Ω × J , u ( x , t ) = 0 , ( x , t ) Ω × J ¯ , u ( x , 0 ) = u 0 ( x ) , x Ω ¯ ,
where Ω R 2 is a convex and bounded polygon region with boundary Ω , J = ( 0 , T ] with 0 < T < . We assume that the source function f ( x , t ) , initial data u 0 ( x ) , and non-negative coefficient p ( x ) are smooth enough. Specifically, for the symmetric diffusion coefficient matrix A ( x ) , we should assume that there exist two constants A 0 , A 1 > 0 such that
A 0 z T z z T A ( x ) z A 1 z T z , z R 2 , x Ω ¯ .
Moreover, the multi-term time-fractional derivative P ( D t ) u ( x , t ) is defined by
P ( D t ) u ( x , t ) = i = 1 m b i D t α i u ( x , t ) , 0 < α m < α m 1 < < α 1 < 1 ,
where b i   ( i = 1 , 2, ⋯, m ) are the positive real numbers and D t α i u is the Caputo time-fractional derivative as follows:
D t α i u ( x , t ) = 1 Γ ( 1 α i ) 0 t u ( x , s ) s 1 ( t s ) α i d s ,
where Γ ( · ) denotes the Γ -function.
It should be noted that the MFE method, as an important numerical calculation method, has been widely used to solve FPDEs [28,29,30,31,32], and some scholars have also used this method to solve the multi-term FPDEs [33,34,35]. In [33], Shi et al. proposed an H 1 -Galerkin mixed finite element (MFE) method for the multi-term time-fractional diffusion equations and gave a superconvergence result. In [34], Li et al. proposed an MFE method for the multi-term time-fractional diffusion and diffusion-wave equations by using an MFE space contained in ( L 2 ( Ω ) ) d × H 0 1 ( Ω ) , where d = 2 , 3 . In [35], Cao et al. constructed a nonconforming MFE scheme for the multi-term time-fractional mixed diffusion and diffusion-wave equations. Motivated by the above excellent works, we will construct a fully discrete MFE scheme for the multi-term TFRD equation (1) by using the Raviart–Thomas MFE space and the L 1 formula, analyze the existence, uniqueness, and unconditional stability in detail, and give error estimates for u (in discrete L ( L 2 ( Ω ) ) norm) and auxiliary variable λ (in discrete L ( ( L 2 ( Ω ) ) 2 ) and discrete L ( H ( div , Ω ) ) norms). Finally, we give numerical experiments to demonstrate the efficiency of the proposed method.
The remainder of this paper is arranged as follows. In Section 2, we construct a fully discrete MFE scheme for the multi-term TFRD equations by using the Raviart–Thomas MFE space and the L 1 -formula. In Section 3, we give a fractional Grönwall inequality and analyze the existence and uniqueness of the discrete solution. We derive the unconditional stability results and a priori error estimates in detail in Section 4 and Section 5, respectively. Finally, three numerical examples are given to verify the theoretical results.

2. Mixed Finite Element Method

We introduce the flux λ ( x , t ) = A ( x ) u ( x , t ) as the auxiliary variable. Then, the original problem (1) can be rewritten as follows:
( a ) P ( D t ) u ( x , t ) + div λ ( x , t ) + p ( x ) u ( x , t ) = f ( x , t ) , ( x , t ) Ω × J , ( b ) A 1 ( x ) λ ( x , t ) + u ( x , t ) = 0 , ( x , t ) Ω × J , ( c ) u ( x , t ) = 0 , ( x , t ) Ω × J ¯ , ( d ) u ( x , 0 ) = u 0 ( x ) , x Ω ¯ .
Let V = L 2 ( Ω ) and W = H ( div , Ω ) = { w ( L 2 ( Ω ) ) 2 : div w L 2 ( Ω ) } . Then, we obtain the mixed variational formulation of (2): find ( u , λ ) V × W such that
( a ) ( P ( D t ) u , v ) + ( div λ , v ) + ( p u , v ) = ( f , v ) , v V , ( b ) ( A 1 λ , w ) ( u , div w ) = 0 , w W , ( c ) u ( x , 0 ) = u 0 ( x ) , x Ω ¯ .
Let K h be a quasi-uniform triangulation of the domain Ω , h T be the diameter of the triangle T K h and denote h = max { h T } . We select the Raviart–Thomas MFE space V h × W h V × W , that is,
V h ( K ) = { v h V : v h | T P r ( T ) , T K h } , W h ( K ) = { w h W : w h | T ( P r ( T ) ) 2 ( x P r ( T ) ) , T K h } ,
where the notation ⊕ indicates a direct sum, x P r ( T ) = ( x 1 P r ( T ) , x 2 P r ( T ) ) , x = ( x 1 , x 2 ) and r 0 is a given integer.
Let τ = T / N and t n = n τ for n = 0 , 1, 2, ⋯, N, where N is a positive integer. For the parameters α i and i = 1 , 2, ⋯, m, the Caputo time-fraction derivative D t α i u ( x , t ) at t = t n is approximated by using the well-known L 1 formula [36,37] as follows:
D t α i u ( x , t n ) = 1 Γ ( 1 α i ) 0 t n u ( x , s ) s 1 ( t n s ) α i d s = 1 Γ ( 2 α i ) k = 0 n 1 u k + 1 u k τ [ ( t n t k ) 1 α i ( t n t k + 1 ) 1 α i ] + Q α i n ( x ) = 1 Γ ( 2 α i ) [ d α i , 1 n u n + k = 1 n 1 ( d α i , k + 1 n d α i , k n ) u n k d α i , n n u 0 ] + Q α i n ( x ) = 1 Γ ( 2 α i ) k = 0 n d ˜ α i , k n u k + Q α i n ( x ) ,
where d α i , k n = ( t n t n k ) 1 α i ( t n t n k + 1 ) 1 α i τ , d ˜ α i , 0 n = d α i , n n , d ˜ α i , n n = d α i , 1 n , and d ˜ α i , k n = d α i , n k + 1 n d α i , n k n ( 0 < k n 1 ) . Setting D N α i u n = 1 Γ ( 2 α i ) k = 0 n d ˜ α i , k n u k , we have D t α i u ( x , t n ) = D N α i u n + Q α i n ( x ) , where Q α i n ( x ) is the truncation error.
Based on the above definitions, and setting u h n and λ h n to be the discrete solutions of u and λ at t = t n , respectively, then we can design a fully discrete MFE scheme for the original problem (1): find ( u h n , λ h n ) V h × W h such that
( a ) ( i = 1 m b i D N α i u h n , v h ) + ( div λ h n , v h ) + ( p u h n , v h ) = ( f n , v h ) , v h V h , ( b ) ( A 1 λ h n , w h ) ( u h n , div w h ) = 0 , w h W h ,
where ( u h 0 , λ h 0 ) V h × W h satisfies
( a ) ( A 1 λ h 0 , w h ) ( u h 0 , div w h ) = 0 , w h W h , ( b ) ( div λ h 0 , v h ) = ( div λ 0 , v h ) , v h V h ,
where λ 0 ( x ) = A ( x ) u 0 ( x ) .
Remark 1.
(I) In the MFE scheme (5)(6), we particularly emphasize the calculation of initial values ( u h 0 , λ h 0 ) , as this calculation will be used in stability and convergence analyses. Moreover, from the mixed elliptic projection R h defined in Section 5, we can see that ( u h 0 , λ h 0 ) = ( R h u 0 , R h λ 0 ) .
(II) Compared with the standard FE methods, it is well known that the MFE method can not only reduce the smoothness requirement of the finite element space, but also simultaneously calculate multiple physical quantities. These advantages are very important and popular in practical applications.

3. Existence and Uniqueness

In this section, we shall prove the existence and uniqueness for the MFE scheme (5)–(6). We first give some lemmas, which are important in subsequent theoretical analysis.
Lemma 1
([38]). There exist two positive constants μ 0 and μ 1 such that
μ 0 w 2 w A 1 2 μ 1 w 2 , where w A 1 2 = ( A 1 w , w ) , w W .
Lemma 2
([28]). Let { z n } n = 0 be a sequence on W h . Then, the following identity holds:
k = 0 n d ˜ α i , k n ( A 1 z k , z n ) = 1 2 [ d ˜ α i , n n ( A 1 z n , z n ) + k = 0 n 1 d ˜ α i , k n ( A 1 z k , z k ) + k = 0 n 1 d ˜ α i , k n ( A 1 ( z n z k ) , z n z k ] .
Lemma 3. 
Let { φ k : 0 k N } be a non-negative sequence, { ξ k : 0 k N } be a nondecreasing positive sequence, and C 0 1 be a constant, which satisfy
i = 1 m b i Γ ( 2 α i ) d ˜ α i , n n φ n C 0 i = 1 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n φ k + ξ n , 1 n N .
Then, we have
φ n C 0 n ( φ 0 + 1 i = 1 m b i Γ ( 2 α i ) d α i , n n ξ n ) , 1 n N .
Further, we can further write (8) as
φ n C 0 n ( φ 0 + i = 1 m Γ ( 1 α i ) t n α i b i ξ n ) , 1 n N .
Proof. 
When n = 1 in (7), we have
i = 1 m b i Γ ( 2 α i ) d ˜ α i , 1 1 φ 1 C 0 i = 1 m b i Γ ( 2 α i ) d ˜ α i , 0 1 φ 0 + ξ 1 .
Noting that d ˜ α i , 0 n = d α i , n n , d ˜ α i , n n = d α i , 1 n , we have
i = 1 m b i Γ ( 2 α i ) d α i , 1 1 φ 1 C 0 i = 1 m b i Γ ( 2 α i ) d α i , 1 1 φ 0 + ξ 1 .
Then, we can obtain
φ 1 C 0 ( φ 0 + 1 i = 1 m b i Γ ( 2 α i ) d α i , 1 1 ξ 1 ) .
It means that the conclusion (8) is valid for the case of n = 1 . Assume that (8) is valid for n = 1 , 2 , , r . We now need to prove that it also holds for n = r + 1 . Selecting n = r + 1 in (7), we obtain
i = 1 m b i Γ ( 2 α i ) d ˜ α i , j + 1 j + 1 φ j + 1 C 0 i = 1 m b i Γ ( 2 α i ) k = 0 j d ˜ α i , k j + 1 φ k + ξ j + 1 = C 0 i = 1 m b i Γ ( 2 α i ) k = 1 j ( d α i , j k + 1 j + 1 d α i , j k + 2 j + 1 ) φ k + C 0 i = 1 m b i d α i , j + 1 j + 1 φ 0 + ξ j + 1 = C 0 i = 1 m b i Γ ( 2 α i ) k = 0 j 1 ( d α i , k + 1 j + 1 d α i , k + 2 j + 1 ) φ j k + C 0 i = 1 m b i Γ ( 2 α i ) d α i , j + 1 j + 1 φ 0 + ξ j + 1 .
Noting that 0 < d α i , k + 1 k + 1 < d α i , k k and 0 < d α i , k + 1 n < d α i , k n , ( k = 0 , 1 j ) , we have
i = 1 m b i Γ ( 2 α i ) d ˜ α i , j + 1 j + 1 φ j + 1 C 0 i = 1 m b i Γ ( 2 α i ) k = 0 j 1 ( d α i , k + 1 j + 1 d α i , k + 2 j + 1 ) [ C 0 j k ( φ 0 + 1 i = 1 m b i Γ ( 2 α i ) d α i , j k j k ξ j k ) ] + C 0 i = 1 m b i Γ ( 2 α i ) d α i , j + 1 j + 1 ( φ 0 + 1 i = 1 m b i Γ ( 2 α i ) d α i , j + 1 j + 1 ξ j + 1 ) C 0 j + 1 i = 1 m b i Γ ( 2 α i ) [ k = 0 j 1 ( d α i , k + 1 j + 1 d α i , k + 2 j + 1 ) + d α i , j + 1 j + 1 ] ( φ 0 + 1 i = 1 m b i Γ ( 2 α i ) d α i , j + 1 j + 1 ξ j + 1 ) C 0 j + 1 i = 1 m b i Γ ( 2 α i ) d α i , 1 j + 1 ( φ 0 + 1 i = 1 m b i Γ ( 2 α i ) d α i , j + 1 j + 1 ξ j + 1 ) .
Therefore, using the mathematical induction method, we can complete the proof of (8). From [37], we know n α i ( n 1 α i ( n 1 ) 1 α i ) 1 α i , and then
d α i , n n = ( n τ ) 1 α i ( n τ τ ) 1 α i τ = ( n 1 α i ( n 1 ) 1 α i ) τ α i ( 1 α i ) τ α i n α i .
Thus, making use of (8) and (15), we can complete the proof of (9). □
Next, we give the existence and uniqueness results for the MFE scheme (5).
Theorem 1. 
The MFE scheme (5) has a unique solution.
Proof. 
Let V h = span { ϕ 1 , ϕ 2 , , ϕ M 1 } and W h = span { ψ 1 , ψ 2 , , ψ M 2 } . Then, u h n and λ h n can be written as
u h n = i = 1 M 1 u ˜ i n ϕ i , λ h n = j = 1 M 2 s ˜ j n ψ j .
Substituting (16) into (5) and selecting v h = ϕ i   ( i = 1 , 2 , , M 1 ) and w h = ψ j   ( j = 1 , 2 , , M 2 ) , we have
i = 1 m b i Γ ( 2 α i ) d ˜ α i , n n B 1 + B 3 D T D B 2 U n L n = F n i = 1 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n B 1 U k 0 ,
where
U n = ( u ˜ 1 n , u ˜ 2 n , , u ˜ M 1 n ) T , L n = ( s ˜ 1 n , s ˜ 2 n , , s ˜ M 2 n ) T , B 1 = ( ( ϕ i , ϕ j ) ) M 1 × M 1 , B 2 = ( ( A 1 ψ i , ψ j ) ) M 2 × M 2 , B 3 = ( ( p ϕ i , ϕ j ) ) M 1 × M 1 , D = ( ( div ψ i , ϕ j ) ) M 2 × M 1 , F n = ( ( f n , ϕ i ) ) M 1 × 1 ,
Noting that B 1 and B 2 are symmetric positive definite matrices and B 3 is a symmetric semi-positive matrix, we have
E D T B 2 1 0 E i = 1 m b i Γ ( 2 α i ) d ˜ α i , n n B 1 + B 3 D T D B 2 = G 0 D B 2 .
where G = i = 1 m b i Γ ( 2 α i ) d ˜ α i , n n B 1 + B 3 + D T B 2 1 D . It is easy to see that G is invertible, so the coefficient matrix of linear Equation (17) is invertible. This means that the MFE scheme (5) has a unique solution. □
Remark 2. 
For Lemma 3, when C 0 = 1 , a similar conclusion can be seen from the proof of Theorem 3.1 in [20]. When C 0 > 1 , some special applications can be seen from [39]. It should be noted that this lemma can be considered a fractional Grönwall inequality without any other conditions for its existence, which will play a crucial role in the subsequent proof process of stability and convergence analyses.

4. Stability Analysis

In this section, we will discuss the unconditional stability for the MFE scheme (5)–(6).
Theorem 2. 
Let ( u h n , λ h n ) n = 1 N be the solutions of the MFE scheme (5). Then, there exists a constant C > 0 independent of h and N such that
u h n u h 0 + i = 1 m Γ ( 1 α i ) t n α i b i sup t [ 0 , T ] f ( t ) U h , λ h n C λ h 0 + i = 1 m Γ ( 1 α i ) t n α i b i 1 / 2 ( sup t [ 0 , T ] f ( t ) + p U h ) .
Proof. 
Taking v h = u h n and w h = λ h n in (5), we have
( i = 1 m b i D N α i u h n , u h n ) + ( A 1 λ h n , λ h n ) + ( p u h n , u h n ) = ( f n , u h n ) .
Using Lemma 1 and the definition of D N α i u h n , we have
i = 1 m b i Γ ( 2 α i ) d ˜ α i , n n ( u h n , u h n ) + μ 0 λ h n 2 i = 1 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n ( u h k , u h n ) + ( f n , u h n ) .
Applying the Cauchy–Schwarz inequality yields
i = 1 m b i Γ ( 2 α i ) d ˜ α i , n n u h n i = 1 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n u h k + f n .
Using Lemma 3, we obtain
u h n u h 0 + i = 1 m Γ ( 1 α i ) t n α i b i sup t [ 0 , T ] f ( t ) U h .
Next, using (5) ( b ) and (6) ( a ) , we have
( A 1 i = 1 m b i D N α i λ h n , w h ) ( i = 1 m b i D N α i u h n , div w h ) = 0 , w h W h .
Choosing v h = i = 1 m b i D N α i u h n and w h = λ h n in (5) ( a ) and (23), respectively, we obtain
i = 1 m b i D N α i u h n 2 + ( A 1 i = 1 m b i D N α i λ h n , λ h n ) + ( p u h n , i = 1 m b i D N α i u h n ) = ( f n , i = 1 m b i D N α i u h n ) .
Using Lemma 2 in (24) yields
i = 1 m b i D N α i u h n 2 + 1 2 i = 0 m b i Γ ( 2 α i ) [ d ˜ α i , n n ( A 1 λ h n , λ h n ) + k = 0 n 1 d ˜ α i , k n ( A 1 λ h k , λ h k ) k = 0 n 1 d ˜ α i , k n ( A 1 ( λ h n λ h k ) , λ h n λ h k ) ] = ( f n , i = 1 m b i D N α i u h n ) ( p u h n , i = 1 m b i D N α i u h n ) .
Because of d ˜ α i , k n < 0 , 0 < k n 1 , we have
i = 1 m b i D N α i u h n 2 + 1 2 i = 0 m b i Γ ( 2 α i ) d ˜ α i , n n ( A 1 λ h n , λ h n ) 1 2 i = 0 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n ( A 1 λ h k , λ h k ) + ( f n , i = 1 m b i D N α i u h n ) ( p u h n , i = 1 m b i D N α i u h n ) .
Apply the Cauchy–Schwarz inequality and the Young inequality in (26) to obtain
i = 1 m b i D N α i u h n 2 + 1 2 i = 0 m b i Γ ( 2 α i ) d ˜ α i , n n ( A 1 λ h n , λ h n ) 1 2 i = 0 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n ( A 1 λ h k , λ h k ) + 1 2 i = 1 m b i D N α i u h n 2 + f n 2 + p 2 u h n 2 .
Using Lemma 3 in (27), we obtain
λ h n C ( λ h 0 + i = 1 m Γ ( 1 α i ) t n α i b i 1 / 2 ( sup t [ 0 , T ] f ( t ) + p U h ) ) .
Thus, we complete the proof. □

5. Convergence Analysis

In this section, we will present the convergence results. For this purpose, we first introduce the mixed elliptic projection ( R h u , R h λ ) V h × W h defined by
( a ) ( A 1 ( λ R h λ ) , w h ) ( u R h u , div w h ) = 0 , w h W h , ( b ) ( div ( λ R h λ ) , v h ) = 0 , v h V h .
Then, the above projection satisfies the classical estimates as follows.
Lemma 4
([40,41]). There exists a constant C > 0 independent of h and N such that
λ R h λ C h r + 1 λ r + 1 , for λ ( H r + 1 ( Ω ) ) 2 , div ( λ R h λ ) C h r + 1 div λ r + 1 , for div λ H r + 1 ( Ω ) , u R h u C h r + 1 ( u r + 1 + λ r + 1 ) , for u H r + 1 ( Ω ) , λ ( H r + 1 ( Ω ) ) 2 .
For the truncation error Q α i n   ( i = 1 , 2 , , m ) of the L 1 formula, from [36,37], we give the following estimates.
Lemma 5. 
Let u C 2 ( J ¯ , L 2 ( Ω ) ) . Then, we have
Q α i n C N ( 2 α i ) , i = 1 , 2 , , m , i = 1 m b i Q α i n C N ( 2 α 1 ) ,
where C > 0 is a constant independent of h and N.
Now, we write the errors u ( t n ) u h n = u ( t n ) R h u ( t n ) + R h u ( t n ) u h n = ρ n + θ n and λ ( t n ) λ h n = λ ( t n ) R h λ ( t n ) + R h λ ( t n ) λ h n = ξ n + η n . From (3) and (5), making use of the mixed elliptic projection R h , we have the following error equations:
( a ) ( i = 1 m b i D N α i ( θ n + ρ n ) , v h ) + ( div η n , v h ) + ( p ( θ n + ρ n ) , v h ) = ( i = 1 m b i Q α i n , v h ) , v h V h , ( b ) ( A 1 η n , w h ) ( θ n , div w h ) = 0 , w h W h .
Noting that ( u h 0 , λ h 0 ) = ( R h u 0 , R h λ 0 ) , we have θ 0 = 0 and η 0 = 0 . We next give the convergence results for the MFE scheme (5)–(6).
Theorem 3. 
Let ( u n , λ n ) V × W and ( u h n , λ h n ) V h × W h be the solutions of (3) and (5), respectively. Assume that u , div λ C 2 ( J ¯ , H r + 1 ( Ω ) ) , λ C 2 ( J ¯ , ( H r + 1 ( Ω ) ) 2 ) . Then, we have
max 1 n N u ( t n ) u h n + max 1 n N λ ( t n ) λ h n C ( h r + 1 + N ( 2 α 1 ) ) , max 1 n N λ ( t n ) λ h n H ( div , Ω ) C ( 1 + N α m 2 ) ( h r + 1 + N ( 2 α 1 ) ) ,
where C > 0 is a constant independent of h and N.
Proof. 
Taking v h = θ n and w h = η n in (30), we can obtain
( i = 1 m b i D N α i θ n , θ n ) + ( A 1 η n , η n ) + ( p θ n , θ n ) = ( i = 1 m b i Q α i n , θ n ) ( p ρ n , θ n ) ( i = 1 m b i D N α i ρ n , θ n ) .
Noting that p ( x ) 0 , using the Lemma 1 and the definition of D N α i u h n , we have
i = 1 m b i Γ ( 2 α i ) d ˜ α i , n n ( θ n , θ n ) + μ 0 η n 2 i = 1 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n ( θ k , θ n ) ( i = 1 m b i Q α i n , θ n ) ( p ρ n , θ n ) ( i = 1 m b i D N α i ρ n , θ n ) .
Applying the Cauchy–Schwarz inequality, we obtain
i = 1 m b i Γ ( 2 α i ) d ˜ α i , n n θ n 2 + μ 0 η n 2 i = 1 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n θ k θ n + ( i = 1 m b i Q α i n + p ρ + i = 1 m b i D N α i ρ n ) θ n ,
and then
i = 1 m b i Γ ( 2 α i ) d ˜ α i , n n θ n 2 i = 1 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n θ k + ( i = 1 m b i Q α i n + p ρ + i = 1 m b i D N α i ρ n ) .
Using Lemmas 4 and 5, we obtain
i = 1 m b i Γ ( 2 α i ) d ˜ α i , n n θ n i = 1 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n θ k + C ( N ( 2 α 1 ) + h r + 1 ) .
Noting that θ 0 = 0 and using Lemma 3, we obtain
θ n C ( h r + 1 + N ( 2 α 1 ) ) .
Now, from (30) ( b ) , we obtain
( A 1 i = 1 m b i D N α i η n , w h ) ( i = 1 m b i D N α i θ n , div w h ) = 0 , w h W h .
Choosing v h = i = 1 m b i D N α i θ n and w h = η n in (30) ( a ) and (37), respectively, we can obtain
i = 1 m b i D N α i θ n 2 + ( A 1 i = 1 m b i D N α i η n , η n ) = ( i = 1 m b i D N α i ρ n , i = 1 m b i D N α i θ n ) ( p ( ρ n + θ n ) , i = 1 m b i D N α i θ n ) ( i = 1 m b i Q α i n , i = 1 m b i D N α i θ n ) .
Using Lemma 2, we have
i = 1 m b i D N α i θ n 2 + 1 2 i = 1 m b i Γ ( 2 α i ) [ d ˜ α i , n n ( A 1 η n , η n ) k = 0 n 1 d ˜ α i , k n ( A 1 ( η n η k , η n η k ) ] = 1 2 i = 1 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n ( A 1 η k , η k ) ( i = 1 m b i D N α i ρ n , i = 1 m b i D N α i θ n ) ( i = 1 m b i Q α i n , i = 1 m b i D N α i θ n ) ( p ( ρ n + θ n ) , i = 1 m b i D N α i θ n ) .
Noting that d ˜ α i , k n < 0 , 0 < k n 1 and using Lemma 1, we obtain
i = 1 m b i D N α i θ n 2 + 1 2 i = 1 m b i Γ ( 2 α i ) d ˜ α i , n n ( A 1 η n , η n ) 1 2 i = 1 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n ( A 1 η k , η k ) ( i = 1 m b i D N α i ρ n , i = 1 m b i D N α i θ n ) ( i = 1 m b i Q α i n , i = 1 m b i D N α i θ n ) ( p ( ρ n + θ n ) , i = 1 m b i D N α i θ n ) .
Applying the Cauchy–Schwarz and the Young inequality in (40) yields
i = 1 m b i D N α i θ n 2 + 1 2 i = 1 m b i Γ ( 2 α i ) d ˜ α i , n n ( A 1 η n , η n ) 1 2 i = 1 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n ( A 1 η k , η k ) + 1 2 i = 1 m b i D N α θ n 2 + 2 ( i = 1 m b i D N α i ρ n + i = 1 m b i Q α i n 2 + p 2 ( ρ n 2 + θ n 2 ) ) .
Using Lemmas 4 and 5, we obtain
i = 1 m b i Γ ( 2 α i ) d ˜ α i , n n ( A 1 η n , η n ) i = 1 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n ( A 1 η k , η k ) + C ( N 2 ( 2 α 1 ) + h 2 r + 2 ) .
Noting that η 0 = 0 and using Lemma 3, we obtain
η n C ( h r + 1 + N ( 2 α 1 ) ) .
We now estimate λ n λ h n H ( div , Ω ) . Taking v h = i = 1 m b i D N α i θ n and w h = η n in (30) ( a ) and (37), respectively, we have
div η n 2 = ( A 1 i = 1 m b i D N α i η n , η n ) ( i = 1 m b i D N α i ρ n , div η n ) ( p ( ρ n + θ n ) , div η n ) ( i = 1 m b i Q α i n , div η n ) .
For the term ( A 1 i = 1 m b i D N α i η n , η n ) , noting that k = 0 n 1 ( d ˜ α i , k n ) = T α i N α i , we obtain
( A 1 i = 1 m b i D N α i η n , η n ) = i = 1 m b i Γ ( 2 α i ) k = 0 n 1 d ˜ α i , k n ( A 1 η k , η n ) + d ˜ α i , n n ( A 1 η n , η n ) μ 1 i = 1 m b i Γ ( 2 α i ) k = 0 n 1 ( d ˜ α i , k n ) η k η n C i = 1 m b i Γ ( 2 α i ) T α i N α i ( N ( 2 α 1 ) + h r + 1 ) 2 .
Then, it holds from (45) that
div η n 2 = 2 ( i = 1 m b i D N α i ρ n + i = 1 m b i Q α i n 2 + p 2 ( ρ n 2 + θ n 2 ) ) + C i = 1 m b i Γ ( 2 α i ) T α i N α i ( N ( 2 α 1 ) + h r + 1 ) 2 + 1 2 div η n 2 .
Using Lemmas 4 and 5, we have
div η n C ( 1 + N α m 2 ) ( h r + 1 + N ( 2 α 1 ) ) .
Then, we finish the proof. □
Remark 3.
(I) For variables u and λ, we define the discrete norms of the errors as follows:
u u h L ^ ( L 2 ( Ω ) ) = max 1 n N u ( t n ) u h n , λ λ h L ^ ( ( L 2 ( Ω ) ) 2 ) = max 1 n N λ ( t n ) λ h n , λ λ h L ^ ( H ( div , Ω ) ) = max 1 n N λ ( t n ) λ h n H ( div , Ω ) .
From Theorem 3, we obtain the optimal a priori error estimate results for u in the discrete L ( L 2 ( Ω ) ) norm and λ in the discrete L ( ( L 2 ( Ω ) ) 2 ) norm and obtain the suboptimal error estimate for λ in the discrete L ( H ( div , Ω ) ) norm. In the actual calculation in the next section, we achieve the optimal convergence rates for variables u and λ based on the above discrete norms.
(II) It should be pointed out that the solutions of many FPDEs have an initial layer at t = 0 (see [42,43]). To overcome this difficulty, some scholars have adopted nonuniform mesh methods and achieved excellent results [24,26,42,44,45,46]. Moreover, it is noted that { ξ k : 0 k N } in Lemma 3 is required to be a nondecreasing positive sequence, so the error estimates for the MFE scheme (5)(6) with the temporal nonuniform method should adopt some other techniques. It is gratifying that the numerical results in Example 3 show that the MFE scheme (5)(6) with the temporal graded mesh is feasible and effective.

6. Numerical Examples

In this section, we given three test examples to verify the effectiveness and convergence accuracy of the proposed MFE scheme (5)–(6) and adopt the lowest-order Raviart–Thomas MFE space for variables u and λ in the numerical experiments.
Example 1. 
Consider the following two-term TFRD equation:
D t α 1 u ( x , t ) + D t α 2 u ( x , t ) Δ u ( x , t ) + p ( x ) u ( x , t ) = f ( x , t ) , ( x , t ) Ω × J , u ( x , t ) = 0 , ( x , t ) Ω × J ¯ , u ( x , 0 ) = u 0 ( x ) , x Ω ¯ ,
where J = ( 0 , 1 ] , Ω = ( 0 , 1 ) 2 , p ( x ) = 1 + x 1 2 + x 2 2 , x = ( x 1 , x 2 ) Ω , u ( x , 0 ) = 0 , and the source function f is taken by
f ( x , t ) = Γ ( 3 + α 1 + α 2 ) Γ ( 3 + α 2 ) t 2 + α 2 + Γ ( 3 + α 1 + α 2 ) Γ ( 3 + α 1 ) t 2 + α 1 + ( 2 π 2 + p ( x ) ) t 2 + α 1 + α 2 × sin ( π x 1 ) sin ( π x 2 ) .
And we can find the analytical solutions for variables u and λ as follows:
u ( x , t ) = t 2 + α 1 + α 2 sin ( π x 1 ) sin ( π x 2 ) , λ ( x , t ) = π t 2 + α 1 + α 2 ( cos ( π x 1 ) sin ( π x 2 ) , sin ( π x 1 ) cos ( π x 2 ) ) .
In the numerical simulation, we select fractional parameters α 1 = 0.9 , 0.7 , 0.5 and α 2 = 0.1, 0.4 in Equation (48) and know that among these different fractional parameters, the convergence rates are only related to the largest fractional parameter α 1 from Theorem 3. By taking N = 5 , 8, 10, 16 and the corresponding h = 2 / N 2 α 1 , we give the error results and convergence rates in Table 1, Table 2 and Table 3 for the MFE scheme (5)–(6), which show that the convergence rates in the temporal direction for u (in the discrete L ( L 2 ( Ω ) ) norm) and λ (in the discrete L ( ( L 2 ( Ω ) ) 2 ) and L ( H ( div , Ω ) ) norms) are close to 2 α 1 . Moreover, in order to test convergence rates in the spatial direction, by fixing N = 100 and taking h = 2 / 4 , 2 / 8 , 2 / 16 , 2 / 32 , we give the error results and convergence rates in Table 4, Table 5 and Table 6, which show that the convergence rates in the spatial direction for u (in the discrete L ( L 2 ( Ω ) ) norm) and λ (in the discrete L ( ( L 2 ( Ω ) ) 2 ) and L ( H ( div , Ω ) ) norms) are close to 1.
Example 2. 
Consider the following three-term TFRD equation:
D t α 1 u ( x , t ) + D t α 2 u ( x , t ) + D t α 3 u ( x , t ) Δ u ( x , t ) + p ( x ) u ( x , t ) = f ( x , t ) , ( x , t ) Ω × J , u   ( x , t ) = 0 , ( x , t ) Ω × J ¯ , u ( x , 0 ) = u 0 ( x ) , x Ω ¯ ,
where the spatial domain Ω, temporal domain J, coefficient p ( x ) , and initial data u ( x , 0 ) are as in Example 1 and the source function f is taken by
f ( x , t ) = ( Γ ( 3 + α 1 + α 2 + α 3 ) Γ ( 3 + α 2 + α 3 ) t 2 + α 2 + α 3 + Γ ( 3 + α 1 + α 2 + α 3 ) Γ ( 3 + α 1 + α 3 ) t 2 + α 1 + α 3 + Γ ( 3 + α 1 + α 2 + α 3 ) Γ ( 3 + α 1 + α 2 ) t 2 + α 1 + α 2 + ( 2 π 2 + p ( x ) ) t 2 + α 1 + α 2 + α 3 ) sin ( π x 1 ) sin ( π x 2 ) .
And we can also find the analytical solutions for variables u and λ as follows:
u ( x , t ) = t 2 + α 1 + α 2 + α 3 sin ( π x 1 ) sin ( π x 2 ) , λ ( x , t ) = π t 2 + α 1 + α 2 + α 3 ( cos ( π x 1 ) sin ( π x 2 ) , sin ( π x 1 ) cos ( π x 2 ) ) .
In this example, since the Equation (49) contains three Caputo time-fractional derivative terms, we specifically take the fractional parameters α 1 = 0.9 , 0.7 , 0.5 and ( α 2 , α 3 ) = ( 0.4 , 0.2 ) , ( 0.3 , 0.1 ) . From Theorem 3, we also point out that the convergence rates are only related to the maximum fractional parameter α 1 . In Table 7, Table 8 and Table 9, for different N = 5 , 8, 10, 16, we give the error results and convergence rates for the MFE scheme (5)–(6), where the spatial grid sizes are also taken as h = 2 / N 2 α 1 . We can also see that the convergence rates in the temporal direction for u (in the discrete L ( L 2 ( Ω ) ) norm) and λ (in the discrete L ( ( L 2 ( Ω ) ) 2 ) and L ( H ( div , Ω ) ) norms) are close to 2 α 1 . Furthermore, in Table 10, Table 11 and Table 12, we also fix N = 100 and take h = 2 / 4 , 2 / 8 , 2 / 16 , 2 / 32 , give the error results and convergence rates, and see that the convergence rates in the spatial direction for u and λ in the above corresponding discrete norms are also close to 1.
Based on the numerical results in Table 1, Table 2, Table 3, Table 4, Table 5, Table 6, Table 7, Table 8, Table 9, Table 10, Table 11 and Table 12 obtained from the above two test examples, we can see that the convergence rates in the spatial and temporal directions for u (in the discrete L ( L 2 ( Ω ) ) norm) and λ (in the discrete L ( ( L 2 ( Ω ) ) 2 ) norm) are in agreement with the theoretical results in Theorem 3, and those for λ (in the discrete L ( H ( div , Ω ) ) norm) are higher than the theoretical result. These results fully demonstrate that the proposed MFE method for the multi-term TFRD equations is effective.
Example 3. 
Consider the two-term TFRD equation in Example 1 with weak regularity solutions near the initial time t = 0 , where the source function f is taken by
f ( x , t ) = ( 2 Γ ( 3 α 1 ) t 2 α 1 + Γ ( 1 + α 1 ) + Γ ( 2 + α 2 ) Γ ( 2 + α 2 α 1 ) t 1 + α 2 α 1 + 2 Γ ( 3 α 2 ) t 2 α 2 + Γ ( 1 + α 1 ) Γ ( 1 + α 1 α 2 ) t α 1 α 2 + Γ ( 2 + α 2 ) t + ( 2 π 2 + p ( x ) ) ( t 2 + t α 1 + t 1 + α 2 ) ) sin ( π x 1 ) sin ( π x 2 ) .
And we can also find the analytical solutions for variables u and λ as follows:
u ( x , t ) = ( t 2 + t α 1 + t 1 + α 2 ) sin ( π x 1 ) sin ( π x 2 ) , λ ( x , t ) = π ( t 2 + t α 1 + t 1 + α 2 ) ( cos ( π x 1 ) sin ( π x 2 ) , sin ( π x 1 ) cos ( π x 2 ) ) .
In this example, we will select the graded mesh to discretize the interval [ 0 , T ] and set t n = T ( n / N ) γ , for n = 0 , 1, 2, ⋯, N, where constant γ 1 is the temporal graded mesh parameter. The ideal optimal error estimates for u (in the discrete L ( L 2 ( Ω ) ) norm) and λ (in the discrete L ( ( L 2 ( Ω ) ) 2 ) and L ( H ( div , Ω ) ) norms) should be O ( N min { γ α 1 , 2 α 1 } + h ) . Here, we will mainly test the convergence rates in the temporal direction with the graded mesh parameter γ = 1 and ( 2 α 1 ) / α 1 . We first conduct numerical experiments with γ = 1 . Then, the optimal convergence rate in the temporal direction is α 1 . For fractional parameters α 1 = 0.9 , 0.7 , 0.5 and α 2 = 0.1 , 0.4 , we take the time mesh parameter N = 20 , 40, 80, 160 and special spatial grid parameters: (i) when α 1 = 0.9 , take h 2 2 / N α 1 ; (ii) when α 1 = 0.7 , take h 2 / N α 1 ; (iii) when α 1 = 0.5 , take h 2 / ( 2 N α 1 ) . Then, we give the numerical results in Table 13, Table 14 and Table 15, which show that the convergence rates in the temporal direction for u (in the discrete L ( L 2 ( Ω ) ) norm) and λ (in the discrete L ( ( L 2 ( Ω ) ) 2 ) and L ( H ( div , Ω ) ) norms) are close to α 1 .
Next, we conduct numerical experiments with γ = ( 2 α 1 ) / α 1 . Then, the optimal convergence rate is 2 α 1 . We take the time mesh parameter N = 5 , 8, 10, 16 and the spatial grid parameter h = 2 / N 2 α 1 . Then, we give the numerical results in Table 16, Table 17 and Table 18 and find that the convergence rates in the temporal direction for u (in the discrete L ( L 2 ( Ω ) ) norm) and λ (in the discrete L ( ( L 2 ( Ω ) ) 2 ) and L ( H ( div , Ω ) ) norms) are close to 2 α 1 . Based on the above discussion, we know that the MFE scheme (5)–(6) with the temporal graded mesh for solving the multi-term TFRD equations with the initial layer is also feasible and effective.

7. Conclusions

This work presents a Raviart–Thomas MFE method for solving the multi-term TFRD equations with variable coefficients by using the well-known L 1 formula. The existence, uniqueness, and unconditional stability of the discrete solution are discussed, and the optimal a priori error estimates for u (in the discrete L ( L 2 ( Ω ) ) norm) and λ (in the discrete L ( ( L 2 ( Ω ) ) 2 ) norm) and the suboptimal a priori error estimate for λ (in the discrete L ( H ( div , Ω ) ) norm) are obtained in this work. In addition, some numerical results are given to demonstrate the effectiveness of the proposed MFE method. In future research, we will try to give theoretical analysis for the MFE method with the temporal graded mesh to solve some FPDEs with the initial layer at t = 0 and apply the MFE method to solve more FPDEs in scientific and engineering fields.

Author Contributions

Conceptualization, J.Z. and S.D.; methodology, Z.F.; software, Z.F.; validation, J.Z., S.D. and Z.F.; formal analysis, J.Z. and S.D.; writing—original draft preparation, J.Z. and S.D.; writing—review and editing, J.Z. and Z.F.; funding acquisition, J.Z. and Z.F. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (11701299), the Scientific Research Projection of Higher Schools of the Inner Mongolia Autonomous Region (NJZY23055), and the Central Government Guided Local Science and Technology Development Fund Project of China (2022ZY0175).

Data Availability Statement

Data are contained within the article.

Acknowledgments

The authors are very grateful to the editors and reviewers for their helpful comments and suggestions on improving this work.

Conflicts of Interest

The authors declares no conflicts of interest.

References

  1. Ross, B. Fractional Calculus and Its Applications; Lecture Notes in Mathematics; Springer: Berlin/Heidelberg, Germany, 1975. [Google Scholar]
  2. Hilfer, R. Applications of Fractional Calculus in Physics; World Scientific: Singapore, 2000. [Google Scholar]
  3. Diethelm, K. The Analysis of Fractional Differential Equations; Springer: Berlin/Heidelberg, Germany, 2010. [Google Scholar]
  4. Hendy, M.H.; Amin, M.M.; Ezzat, M.A. Two-dimensional problem for thermoviscoelastic materials with fractional order heat transfer. J. Therm. Stress. 2019, 42, 1298–1315. [Google Scholar] [CrossRef]
  5. Zhou, J.C.; Salahshour, S.; Ahmadian, A.; Senu, N. Modeling the dynamics of COVID-19 using fractal-fractional operator with a case study. Results Phys. 2022, 33, 105103. [Google Scholar] [CrossRef] [PubMed]
  6. Shymanskyi, V.; Sokolovskyy, I.; Sokolovskyy, Y.; Bubnyak, T. Variational Method for Solving the Time-Fractal Heat Conduction Problem in the Claydite-Block Construction. In Advances in Computer Science for Engineering and Education. ICCSEEA 2022; Hu, Z., Dychka, I., Petoukhov, S., He, M., Eds.; Lecture Notes on Data Engineering and Communications Technologies; Springer: Cham, Switzerland, 2022; Volume 134. [Google Scholar]
  7. Bagley, R.L.; Torvik, R.J. On the appearance of the fractional derivative in the behaviour of real materials. Appl. Mech. 1984, 51, 294–298. [Google Scholar]
  8. Ming, C.Y.; Liu, F.W.; Zheng, L.C.; Turner, I.; Anh, V. Analytical solutions of multi-term time fractional differential equations and application to unsteady flows of generalized viscoelastic fluid. Comput. Math. Appl. 2006, 72, 2084–2097. [Google Scholar] [CrossRef]
  9. Fetecau, C.; Athar, M.; Fetecau, C. Unsteady flow of a generalized Maxwell fluid with fractional derivative due to a constantly accelerating plate. Comput. Math. Appl. 2009, 57, 596–603. [Google Scholar] [CrossRef]
  10. Liu, F.W.; Meerschaert, M.M.; McGough, R.J.; Zhuang, P.H.; Liu, Q.X. Numerical methods for solving the multi-term time-fractional wave-diffusion equation. Fract. Calc. Appl. Anal. 2013, 16, 9–25. [Google Scholar] [CrossRef]
  11. Dehghan, M.; Safarpoor, M.; Abbaszadeh, M. Two high-order numerical algorithms for solving the multi-term time fractional diffusion-wave equations. J. Comput. Appl. Math. 2015, 290, 174–195. [Google Scholar] [CrossRef]
  12. Ren, J.C.; Sun, Z.Z. Efficient Numerical Solution of the Multi-Term Time Fractional Diffusion-Wave Equation. East Asian J. Appl. Math. 2015, 5, 1–28. [Google Scholar] [CrossRef]
  13. Zheng, M.; Liu, F.W.; Anh, V.; Turner, I. A high-order spectral method for the multi-term time-fractional diffusion equations. Appl. Math. Model. 2016, 40, 4970–4985. [Google Scholar] [CrossRef]
  14. Du, R.L.; Sun, Z.Z. Temporal second-order difference methods for solving multi-term time fractional mixed diffusion and wave equations. Numer. Algorithms 2021, 88, 191–226. [Google Scholar] [CrossRef]
  15. Hendy, A.S.; Zaky, M.A. Graded mesh discretization for coupled system of nonlinear multi-term time-space fractional diffusion equations. Eng. Comput. 2022, 38, 1351–1363. [Google Scholar] [CrossRef]
  16. Liu, Y.Q.; Yin, X.L.; Liu, F.W.; Xin, X.Y.; Shen, Y.F.; Feng, L.B. An alternating direction implicit Legendre spectral method for simulating a 2D multi-term time-fractional Oldroyd-B fluid type diffusion equation. Comput. Math. Appl. 2022, 113, 160–173. [Google Scholar] [CrossRef]
  17. Wei, L.L.; Wang, H.H. Local discontinuous Galerkin method for multi-term variable-order time fractional diffusion equation. Math. Comput. Simulat. 2023, 203, 685–698. [Google Scholar] [CrossRef]
  18. She, M.F.; Li, D.F.; Sun, H.W. A transformed L1 method for solving the multi-term time-fractional diffusion problem. Math. Comput. Simulat. 2022, 193, 584–606. [Google Scholar] [CrossRef]
  19. Jin, B.T.; Lazarov, R.; Liu, Y.K.; Zhou, Z. The Galerkin finite element method for a multi-term time-fractional diffusion equation. J. Comput. Phys. 2015, 281, 825–843. [Google Scholar] [CrossRef]
  20. Li, M.; Huang, C.M.; Jiang, F.Z. Galerkin finite element method for higher dimensional multi-term fractional diffusion equation on non-uniform meshes. Appl. Anal. 2017, 96, 1269–1284. [Google Scholar] [CrossRef]
  21. Zhou, J.; Xu, D.; Chen, H.B. A Weak Galerkin Finite Element Method for Multi-Term Time-Fractional Diffusion Equations. East Asian J. Appl. Math. 2018, 8, 181–193. [Google Scholar] [CrossRef]
  22. Bu, W.P.; Shu, S.; Xue, X.Q.; Xiao, A.G.; Zeng, W. Space-time finite element method for the multi-term time-space fractional diffusion equation on a two-dimensional domain. Comput. Math. Appl. 2019, 78, 1367–1379. [Google Scholar] [CrossRef]
  23. Feng, L.B.; Liu, F.W.; Turner, I. Finite difference/finite element method for a novel 2D multi-term time-fractional mixed sub-diffusion and diffusion-wave equation on convex domains. Commun. Nonlinear Sci. Numer. Simulat. 2019, 70, 354–371. [Google Scholar] [CrossRef]
  24. Meng, X.Y.; Stynes, M. Barrier function local and global analysis of an L1 finite element method for a multiterm time-fractional initial-boundary value problem. J. Sci. Comput. 2020, 84, 5. [Google Scholar] [CrossRef]
  25. Yin, B.L.; Liu, Y.; Li, H.; Zeng, F.H. A class of efficient time-stepping methods for multi-term time-fractional reaction-diffusion-wave equations. Appl. Numer. Math. 2021, 165, 56–82. [Google Scholar] [CrossRef]
  26. Huang, C.B.; Stynes, M.; Chen, H. An α-robust finite element method for a multi-term time-fractional diffusion problem. J. Comput. Appl. Math. 2021, 389, 113334. [Google Scholar] [CrossRef]
  27. Liu, H.; Zheng, X.C.; Fu, H.F. Analysis of a multi-term variable-order time-fractional diffusion equation and its Galerkin finite element approximation. J. Comput. Math. 2022, 40, 814–834. [Google Scholar] [CrossRef]
  28. Zhao, Y.M.; Chen, P.; Bu, W.P.; Liu, X.T.; Tang, Y.F. Two mixed finite element methods for time-fractional diffusion equations. J. Sci. Comput. 2017, 70, 407–428. [Google Scholar] [CrossRef]
  29. Liu, Y.; Fang, Z.C.; Li, H.; He, S. A mixed finite element method for a time-fractional fourth-order partial differential equation. Appl. Math. Comput. 2014, 243, 703–717. [Google Scholar] [CrossRef]
  30. Li, X.C.; Yang, X.Y.; Zhang, Y.H. Error estimates of mixed finite element methods for time-fractional Navier-Stokes equations. J. Sci. Comput. 2017, 70, 500–515. [Google Scholar] [CrossRef]
  31. Liu, Y.; Du, Y.W.; Li, H.; Li, J.C.; He, S. A two-grid mixed finite element method for a nonlinear fourth-order reaction-diffusion problem with time-fractional derivative. Comput. Math. Appl. 2015, 70, 2474–2492. [Google Scholar] [CrossRef]
  32. Li, X.W.; Tang, Y.L. Interpolated coefficient mixed finite elements for semilinear time fractional diffusion equations. Fractal Fract. 2023, 7, 482. [Google Scholar] [CrossRef]
  33. Shi, Z.G.; Zhao, Y.M.; Tang, Y.F.; Wang, F.L.; Shi, Y.H. Superconvergence analysis of an H1-Galerkin mixed finite element method for two-dimensional multi-term time fractional diffusion equations. Int. J. Comput. Math. 2018, 95, 1845–1857. [Google Scholar] [CrossRef]
  34. Li, M.; Huang, C.M.; Ming, W.Y. Mixed finite-element method for multi-term time-fractional diffusion and diffusion-wave equations. Comp. Appl. Math. 2018, 37, 2309–2334. [Google Scholar] [CrossRef]
  35. Cao, F.F.; Zhao, Y.M.; Wang, F.L.; Shi, Y.H.; Yao, C.H. Nonconforming mixed FEM analysis for multi-term time-fractional mixed sub-diffusion and diffusion-wave equation with time-space coupled derivative. Adv. Appl. Math. Mech. 2023, 15, 322–358. [Google Scholar] [CrossRef]
  36. Sun, Z.Z.; Wu, X.N. A fully discrete scheme for a diffusion-wave system. Appl. Numer. Math. 2006, 56, 193–209. [Google Scholar] [CrossRef]
  37. Lin, Y.M.; Xu, C.J. Finite difference/spectral approximations for the time-fractional diffusion equation. J. Comput. Phys. 2007, 225, 1533–1552. [Google Scholar] [CrossRef]
  38. Sinha, R.K.; Ewing, R.E.; Lazarov, R.D. Mixed finite element approximations of parabolic integro-differential equations with nonsmooth initial data. SIAM J. Numer. Anal. 2009, 47, 3269–3292. [Google Scholar] [CrossRef]
  39. Fang, Z.C.; Zhao, J.; Li, H.; Liu, Y. Finite volume element methods for two-dimensional time fractional reaction-diffusion equations on triangular grids. Appl. Anal. 2023, 102, 2248–2270. [Google Scholar] [CrossRef]
  40. Douglas, J.; Roberts, J.E. Global Estimates for Mixed Methods for Second Order Elliptic Equations. Math. Comput. 1985, 44, 39–52. [Google Scholar] [CrossRef]
  41. Durán, R.G. Mixed Finite Element Methods; Lecture Notes in Mathematics; Springer: Berlin/Heidelberg, Germany, 2008; Volume 1939. [Google Scholar]
  42. Stynes, M.; O’Riordan, E.; Gracia, J.L. Error analysis of a finite difference method on graded meshes for a time-fractional diffusion equation. SIAM J. Numer. Anal. 2017, 55, 1057–1079. [Google Scholar] [CrossRef]
  43. Jin, B.T.; Li, B.Y.; Zhou, Z. Numerical analysis of nonlinear subdiffusion equations. SIAM J. Numer. Anal. 2018, 56, 1–23. [Google Scholar] [CrossRef]
  44. Liao, H.L.; Li, D.F.; Zhang, J.W. Sharp error estimate of the nonuniform L1 formula for linear reaction-subdiffusion equations. SIAM J. Numer. Anal. 2018, 56, 1112–1133. [Google Scholar] [CrossRef]
  45. Ren, J.C.; Liao, H.L.; Zhang, Z.M. Superconvergence error estimate of a finite element method on nonuniform time meshes for reaction-subdiffusion equations. J. Sci. Comput. 2020, 84, 38. [Google Scholar] [CrossRef]
  46. Huang, C.B.; Stynes, M. Optimal H1 spatial convergence of a fully discrete finite element method for the time-fractional Allen-Cahn equation. Adv. Comput. Math. 2020, 46, 63. [Google Scholar] [CrossRef]
Table 1. Numerical results with h 2 / N 2 α 1 and α 1 = 0.9 in Example 1.
Table 1. Numerical results with h 2 / N 2 α 1 and α 1 = 0.9 in Example 1.
α 2 N u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.15 8.6930 × 10 2 3.3630 × 10 1 1.7545 × 10 + 0
8 5.2424 × 10 2 1.0760 2.0229 × 10 1 1.0814 1.0569 × 10 + 0 1.0784
10 4.0389 × 10 2 1.1687 1.5578 × 10 1 1.1709 8.1390 × 10 1 1.1708
16 2.3908 × 10 2 1.1157 9.2174 × 10 2 1.1165 4.8141 × 10 1 1.1172
0.45 8.7218 × 10 2 3.3776 × 10 1 1.7631 × 10 + 0
8 5.2624 × 10 2 1.0750 2.0332 × 10 1 1.0799 1.0619 × 10 + 0 1.0786
10 4.0555 × 10 2 1.1674 1.5663 × 10 1 1.1692 8.1786 × 10 1 1.1704
16 2.4011 × 10 2 1.1153 9.2701 × 10 2 1.1160 4.8371 × 10 1 1.1175
Table 2. Numerical results with h 2 / N 2 α 1 and α 1 = 0.7 in Example 1.
Table 2. Numerical results with h 2 / N 2 α 1 and α 1 = 0.7 in Example 1.
α 2 N u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.15 6.5262 × 10 2 2.5184 × 10 1 1.3146 × 10 + 0
8 3.4916 × 10 2 1.3308 1.3451 × 10 1 1.3344 7.0291 × 10 1 1.3321
10 2.6204 × 10 2 1.2863 1.0092 × 10 1 1.2875 5.2740 × 10 1 1.2873
16 1.4175 × 10 2 1.3073 5.4581 × 10 2 1.3077 2.8520 × 10 1 1.3080
0.45 6.5379 × 10 2 2.5244 × 10 1 1.3184 × 10 + 0
8 3.4998 × 10 2 1.3296 1.3493 × 10 1 1.3328 7.0499 × 10 1 1.3318
10 2.6269 × 10 2 1.2858 1.0125 × 10 1 1.2869 5.2895 × 10 1 1.2875
16 1.4212 × 10 2 1.3070 5.4771 × 10 2 1.3073 2.8602 × 10 1 1.3081
Table 3. Numerical results with h 2 / N 2 α 1 and α 1 = 0.5 in Example 1.
Table 3. Numerical results with h 2 / N 2 α 1 and α 1 = 0.5 in Example 1.
α 2 N u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.15 4.7518 × 10 2 1.8311 × 10 1 9.5625 × 10 1
8 2.2765 × 10 2 1.5657 8.7633 × 10 2 1.5679 4.5800 × 10 1 1.5663
10 1.6367 × 10 2 1.4786 6.2996 × 10 2 1.4792 3.2925 × 10 1 1.4791
16 8.1859 × 10 3 1.4742 3.1504 × 10 2 1.4743 1.6465 × 10 1 1.4744
0.45 4.7568 × 10 2 1.8337 × 10 1 9.5799 × 10 1
8 2.2802 × 10 2 1.5645 8.7824 × 10 2 1.5662 4.5893 × 10 1 1.5658
10 1.6396 × 10 2 1.4781 6.3145 × 10 2 1.4785 3.2993 × 10 1 1.4790
16 8.2016 × 10 3 1.4738 3.1585 × 10 2 1.4739 1.6499 × 10 1 1.4744
Table 4. Numerical results with τ = T / N = 1 / 100 and α 1 = 0.9 in Example 1.
Table 4. Numerical results with τ = T / N = 1 / 100 and α 1 = 0.9 in Example 1.
α 2 h u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.1 2 / 4 1.2927 × 10 1 5.0254 × 10 1 2.5970 × 10 + 0
2 / 8 6.5234 × 10 2 0.9867 2.5171 × 10 1 0.9975 1.3118 × 10 + 0 0.9853
2 / 16 3.2696 × 10 2 0.9965 1.2589 × 10 1 0.9996 6.5762 × 10 1 0.9963
2 / 32 1.6361 × 10 2 0.9989 6.2964 × 10 2 0.9996 3.2908 × 10 1 0.9988
0.4 2 / 4 1.2926 × 10 1 5.0244 × 10 1 2.5971 × 10 + 0
2 / 8 6.5231 × 10 2 0.9866 2.5169 × 10 1 0.9973 1.3119 × 10 + 0 0.9853
2 / 16 3.2696 × 10 2 0.9964 1.2589 × 10 1 0.9995 6.5766 × 10 1 0.9962
2 / 32 1.6363 × 10 2 0.9987 6.2973 × 10 2 0.9994 3.2913 × 10 1 0.9987
Table 5. Numerical results with τ = T / N = 1 / 100 and α 1 = 0.7 in Example 1.
Table 5. Numerical results with τ = T / N = 1 / 100 and α 1 = 0.7 in Example 1.
α 2 h u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.1 2 / 4 1.2929 × 10 1 5.0266 × 10 1 2.5969 × 10 + 0
2 / 8 6.5241 × 10 2 0.9867 2.5174 × 10 1 0.9976 1.3118 × 10 + 0 0.9853
2 / 16 3.2698 × 10 2 0.9966 1.2590 × 10 1 0.9997 6.5757 × 10 1 0.9963
2 / 32 1.6359 × 10 2 0.9991 6.2954 × 10 2 0.9999 3.2900 × 10 1 0.9991
0.4 2 / 4 1.2928 × 10 1 5.0258 × 10 1 2.5970 × 10 + 0
2 / 8 6.5239 × 10 2 0.9867 2.5173 × 10 1 0.9975 1.3118 × 10 + 0 0.9853
2 / 16 3.2697 × 10 2 0.9966 1.2590 × 10 1 0.9996 6.5758 × 10 1 0.9963
2 / 32 1.6359 × 10 2 0.9991 6.2954 × 10 2 0.9999 3.2900 × 10 1 0.9990
Table 6. Numerical results with τ = T / N = 1 / 100 and α 1 = 0.5 in Example 1.
Table 6. Numerical results with τ = T / N = 1 / 100 and α 1 = 0.5 in Example 1.
α 2 h u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.1 2 / 4 1.2930 × 10 1 5.0273 × 10 1 2.5969 × 10 + 0
2 / 8 6.5243 × 10 2 0.9868 2.5176 × 10 1 0.9977 1.3118 × 10 + 0 0.9852
2 / 16 3.2699 × 10 2 0.9966 1.2591 × 10 1 0.9997 6.5756 × 10 1 0.9963
2 / 32 1.6359 × 10 2 0.9991 6.2955 × 10 2 0.9999 3.2899 × 10 1 0.9991
0.4 2 / 4 1.2929 × 10 1 5.0266 × 10 1 2.5969 × 10 + 0
2 / 8 6.5242 × 10 2 0.9867 2.5175 × 10 1 0.9976 1.3118 × 10 + 0 0.9853
2 / 16 3.2698 × 10 2 0.9966 1.2590 × 10 1 0.9997 6.5757 × 10 1 0.9963
2 / 32 1.6359 × 10 2 0.9991 6.2955 × 10 2 0.9999 3.2899 × 10 1 0.9991
Table 7. Numerical results with h 2 / N 2 α 1 and α 1 = 0.9 in Example 2.
Table 7. Numerical results with h 2 / N 2 α 1 and α 1 = 0.9 in Example 2.
α 2 α 3 N u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.40.25 8.7400 × 10 2 3.3868 × 10 1 1.7681 × 10 + 0
8 5.2741 × 10 2 1.0747 2.0392 × 10 1 1.0795 1.0648 × 10 + 0 1.0790
10 4.0650 × 10 2 1.1670 1.5711 × 10 1 1.1686 8.2004 × 10 1 1.1704
16 2.4067 × 10 2 1.1152 9.2988 × 10 2 1.1159 4.8494 × 10 1 1.1177
0.30.15 8.7138 × 10 2 3.3735 × 10 1 1.7608 × 10 + 0
8 5.2567 × 10 2 1.0753 2.0303 × 10 1 1.0804 1.0605 × 10 + 0 1.0788
10 4.0507 × 10 2 1.1679 1.5638 × 10 1 1.1698 8.1673 × 10 1 1.1706
16 2.3980 × 10 2 1.1154 9.2546 × 10 2 1.1162 4.8304 × 10 1 1.1175
Table 8. Numerical results with h 2 / N 2 α 1 and α 1 = 0.7 in Example 2.
Table 8. Numerical results with h 2 / N 2 α 1 and α 1 = 0.7 in Example 2.
α 2 α 3 N u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.40.25 6.5465 × 10 2 2.5288 × 10 1 1.3208 × 10 + 0
8 3.5052 × 10 2 1.3291 1.3521 × 10 1 1.3321 7.0629 × 10 1 1.3319
10 2.6310 × 10 2 1.2857 1.0146 × 10 1 1.2868 5.2989 × 10 1 1.2877
16 1.4234 × 10 2 1.3070 5.4886 × 10 2 1.3073 2.8651 × 10 1 1.3083
0.30.15 6.5343 × 10 2 2.5225 × 10 1 1.3173 × 10 + 0
8 3.4972 × 10 2 1.3300 1.3479 × 10 1 1.3334 7.0434 × 10 1 1.3321
10 2.6247 × 10 2 1.2860 1.0114 × 10 1 1.2872 5.2845 × 10 1 1.2876
16 1.4200 × 10 2 1.3071 5.4707 × 10 2 1.3075 2.8575 × 10 1 1.3082
Table 9. Numerical results with h 2 / N 2 α 1 and α 1 = 0.5 in Example 2.
Table 9. Numerical results with h 2 / N 2 α 1 and α 1 = 0.5 in Example 2.
α 2 α 3 N u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.40.25 4.7614 × 10 2 1.8360 × 10 1 9.5930 × 10 1
8 2.2831 × 10 2 1.5638 8.7972 × 10 2 1.5654 4.5961 × 10 1 1.5656
10 1.6417 × 10 2 1.4779 6.3255 × 10 2 1.4782 3.3041 × 10 1 1.4790
16 8.2126 × 10 3 1.4738 3.1641 × 10 2 1.4738 1.6523 × 10 1 1.4745
0.30.15 4.7550 × 10 2 1.8327 × 10 1 9.5743 × 10 1
8 2.2788 × 10 2 1.5650 8.7752 × 10 2 1.5669 4.5859 × 10 1 1.5661
10 1.6385 × 10 2 1.4783 6.3087 × 10 2 1.4788 3.2967 × 10 1 1.4791
16 8.1952 × 10 3 1.4740 3.1552 × 10 2 1.4742 1.6486 × 10 1 1.4745
Table 10. Numerical results with τ = T / N = 1 / 100 and α 1 = 0.9 in Example 2.
Table 10. Numerical results with τ = T / N = 1 / 100 and α 1 = 0.9 in Example 2.
α 2 α 3 h u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.40.2 2 / 4 1.2924 × 10 1 5.0230 × 10 1 2.5973 × 10 + 0
2 / 8 6.5229 × 10 2 0.9865 2.5168 × 10 1 0.9970 1.3119 × 10 + 0 0.9853
2 / 16 3.2696 × 10 2 0.9964 1.2589 × 10 1 0.9994 6.5768 × 10 1 0.9962
2 / 32 1.6364 × 10 2 0.9986 6.2979 × 10 2 0.9992 3.2916 × 10 1 0.9986
0.30.1 2 / 4 1.2925 × 10 1 5.0236 × 10 1 2.5972 × 10 + 0
2 / 8 6.5231 × 10 2 0.9865 2.5169 × 10 1 0.9971 1.3119 × 10 + 0 0.9853
2 / 16 3.2696 × 10 2 0.9964 1.2589 × 10 1 0.9995 6.5765 × 10 1 0.9962
2 / 32 1.6362 × 10 2 0.9988 6.2971 × 10 2 0.9994 3.2912 × 10 1 0.9987
Table 11. Numerical results with τ = T / N = 1 / 100 and α 1 = 0.7 in Example 2.
Table 11. Numerical results with τ = T / N = 1 / 100 and α 1 = 0.7 in Example 2.
α 2 α 3 h u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.40.2 2 / 4 1.2926 × 10 1 5.0244 × 10 1 2.5971 × 10 + 0
2 / 8 6.5237 × 10 2 0.9866 2.5172 × 10 1 0.9972 1.3118 × 10 + 0 0.9853
2 / 16 3.2697 × 10 2 0.9965 1.2590 × 10 1 0.9996 6.5758 × 10 1 0.9963
2 / 32 1.6359 × 10 2 0.9991 6.2954 × 10 2 0.9999 3.2901 × 10 1 0.9990
0.30.1 2 / 4 1.2927 × 10 1 5.0249 × 10 1 2.5970 × 10 + 0
2 / 8 6.5238 × 10 2 0.9866 2.5172 × 10 1 0.9973 1.3118 × 10 + 0 0.9853
2 / 16 3.2697 × 10 2 0.9965 1.2590 × 10 1 0.9996 6.5758 × 10 1 0.9963
2 / 32 1.6359 × 10 2 0.9991 6.2954 × 10 2 0.9999 3.2900 × 10 1 0.9991
Table 12. Numerical results with τ = T / N = 1 / 100 and α 1 = 0.5 in Example 2.
Table 12. Numerical results with τ = T / N = 1 / 100 and α 1 = 0.5 in Example 2.
α 2 α 3 h u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.40.2 2 / 4 1.2927 × 10 1 5.0252 × 10 1 2.5970 × 10 + 0
2 / 8 6.5240 × 10 2 0.9866 2.5173 × 10 1 0.9973 1.3118 × 10 + 0 0.9853
2 / 16 3.2698 × 10 2 0.9965 1.2590 × 10 1 0.9996 6.5757 × 10 1 0.9963
2 / 32 1.6359 × 10 2 0.9991 6.2955 × 10 2 0.9999 3.2899 × 10 1 0.9991
0.30.1 2 / 4 1.2928 × 10 1 5.0257 × 10 1 2.5970 × 10 + 0
2 / 8 6.5241 × 10 2 0.9866 2.5174 × 10 1 0.9974 1.3118 × 10 + 0 0.9853
2 / 16 3.2698 × 10 2 0.9966 1.2590 × 10 1 0.9996 6.5757 × 10 1 0.9963
2 / 32 1.6359 × 10 2 0.9991 6.2955 × 10 2 0.9999 3.2899 × 10 1 0.9991
Table 13. Numerical results with α 1 = 0.9 and graded mesh parameter γ = 1 in Example 3.
Table 13. Numerical results with α 1 = 0.9 and graded mesh parameter γ = 1 in Example 3.
α 2 N u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.120 1.9572 × 10 1 7.5520 × 10 1 3.9354 × 10 + 0
40 1.1208 × 10 1 0.8042 4.3165 × 10 1 0.8070 2.2539 × 10 + 0 0.8041
80 6.0396 × 10 2 0.8920 2.3245 × 10 1 0.8930 1.2146 × 10 + 0 0.8920
160 3.2722 × 10 2 0.8842 1.2591 × 10 1 0.8845 6.5806 × 10 1 0.8842
0.420 1.9571 × 10 1 7.5516 × 10 1 3.9354 × 10 + 0
40 1.1208 × 10 1 0.8042 4.3164 × 10 1 0.8069 2.2540 × 10 + 0 0.8041
80 6.0396 × 10 2 0.8920 2.3244 × 10 1 0.8929 1.2146 × 10 + 0 0.8920
160 3.2722 × 10 2 0.8842 1.2591 × 10 1 0.8845 6.5806 × 10 1 0.8842
Table 14. Numerical results with α 1 = 0.7 and graded mesh parameter γ = 1 in Example 3.
Table 14. Numerical results with α 1 = 0.7 and graded mesh parameter γ = 1 in Example 3.
α 2 N u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.120 1.9573 × 10 1 7.5526 × 10 1 3.9353 × 10 + 0
40 1.2068 × 10 1 0.6976 4.6487 × 10 1 0.7002 2.4269 × 10 + 0 0.6974
80 7.4765 × 10 2 0.6908 2.8779 × 10 1 0.6918 1.5035 × 10 + 0 0.6907
160 4.4872 × 10 2 0.7365 1.7268 × 10 1 0.7369 9.0241 × 10 1 0.7365
0.420 1.9572 × 10 1 7.5523 × 10 1 3.9354 × 10 + 0
40 1.2068 × 10 1 0.6976 4.6486 × 10 1 0.7001 2.4269 × 10 + 0 0.6974
80 7.4765 × 10 2 0.6908 2.8779 × 10 1 0.6918 1.5035 × 10 + 0 0.6907
160 4.4872 × 10 2 0.7365 1.7268 × 10 1 0.7369 9.0241 × 10 1 0.7365
Table 15. Numerical results with α 1 = 0.5 and graded mesh parameter γ = 1 in Example 3.
Table 15. Numerical results with α 1 = 0.5 and graded mesh parameter γ = 1 in Example 3.
α 2 N u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.120 1.7410 × 10 1 6.7141 × 10 1 3.5006 × 10 + 0
40 1.2069 × 10 1 0.5286 4.6487 × 10 1 0.5303 2.4269 × 10 + 0 0.5285
80 8.7212 × 10 2 0.4687 3.3576 × 10 1 0.4694 1.7538 × 10 + 0 0.4686
160 6.2812 × 10 2 0.4735 2.4175 × 10 1 0.4739 1.2632 × 10 + 0 0.4735
0.420 1.7410 × 10 1 6.7139 × 10 1 3.5006 × 10 + 0
40 1.2069 × 10 1 0.5286 4.6487 × 10 1 0.5303 2.4269 × 10 + 0 0.5285
80 8.7212 × 10 2 0.4687 3.3576 × 10 1 0.4694 1.7538 × 10 + 0 0.4686
160 6.2812 × 10 2 0.4735 2.4175 × 10 1 0.4739 1.2632 × 10 + 0 0.4735
Table 16. Numerical results with α 1 = 0.9 and graded mesh parameter γ = 2 α 1 α 1 in Example 3.
Table 16. Numerical results with α 1 = 0.9 and graded mesh parameter γ = 2 α 1 α 1 in Example 3.
α 2 N u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.15 2.6024 × 10 1 1.0062 × 10 + 0 5.2342 × 10 + 0
8 1.5673 × 10 1 1.0789 6.0412 × 10 1 1.0855 3.1526 × 10 + 0 1.0787
10 1.2067 × 10 1 1.1718 4.6479 × 10 1 1.1750 2.4273 × 10 + 0 1.1718
16 7.1368 × 10 2 1.1175 2.7470 × 10 1 1.1189 1.4355 × 10 + 0 1.1176
0.45 2.6021 × 10 1 1.0060 × 10 + 0 5.2351 × 10 + 0
8 1.5673 × 10 1 1.0787 6.0410 × 10 1 1.0852 3.1531 × 10 + 0 1.0787
10 1.2067 × 10 1 1.1716 4.6479 × 10 1 1.1748 2.4276 × 10 + 0 1.1718
16 7.1372 × 10 2 1.1174 2.7472 × 10 1 1.1188 1.4357 × 10 + 0 1.1176
Table 17. Numerical results with α 1 = 0.7 and graded mesh parameter γ = 2 α 1 α 1 in Example 3.
Table 17. Numerical results with α 1 = 0.7 and graded mesh parameter γ = 2 α 1 α 1 in Example 3.
α 2 N u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.15 1.9568 × 10 1 7.5499 × 10 1 3.9360 × 10 + 0
8 1.0462 × 10 1 1.3323 4.0284 × 10 1 1.3365 2.1043 × 10 + 0 1.3323
10 7.8497 × 10 2 1.2872 3.0217 × 10 1 1.2887 1.5789 × 10 + 0 1.2873
16 4.2450 × 10 2 1.3079 1.6336 × 10 1 1.3086 8.5380 × 10 1 1.3081
0.45 1.9566 × 10 1 7.5492 × 10 1 3.9370 × 10 + 0
8 1.0462 × 10 1 1.3320 4.0289 × 10 1 1.3361 2.1049 × 10 + 0 1.3323
10 7.8506 × 10 2 1.2870 3.0221 × 10 1 1.2885 1.5793 × 10 + 0 1.2874
16 4.2457 × 10 2 1.3078 1.6339 × 10 1 1.3084 8.5400 × 10 1 1.3081
Table 18. Numerical results with α 1 = 0.5 and graded mesh parameter γ = 2 α 1 α 1 in Example 3.
Table 18. Numerical results with α 1 = 0.5 and graded mesh parameter γ = 2 α 1 α 1 in Example 3.
α 2 N u L ^ ( L 2 ) Rates λ L ^ ( ( L 2 ) 2 ) Rates λ L ^ ( H ( div ) ) Rates
0.15 1.4253 × 10 1 5.4923 × 10 1 2.8673 × 10 + 0
8 6.8272 × 10 2 1.5661 2.6278 × 10 1 1.5685 1.3733 × 10 + 0 1.5663
10 4.9083 × 10 2 1.4788 1.8890 × 10 1 1.4794 9.8727 × 10 1 1.4790
16 2.4548 × 10 2 1.4742 9.4464 × 10 2 1.4744 4.9373 × 10 1 1.4744
0.45 1.4255 × 10 1 5.4933 × 10 1 2.8688 × 10 + 0
8 6.8306 × 10 2 1.5654 2.6296 × 10 1 1.5675 1.3743 × 10 + 0 1.5659
10 4.9113 × 10 2 1.4783 1.8905 × 10 1 1.4788 9.8804 × 10 1 1.4787
16 2.4567 × 10 2 1.4739 9.4561 × 10 2 1.4740 4.9416 × 10 1 1.4742
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Zhao, J.; Dong, S.; Fang, Z. A Mixed Finite Element Method for the Multi-Term Time-Fractional Reaction–Diffusion Equations. Fractal Fract. 2024, 8, 51. https://doi.org/10.3390/fractalfract8010051

AMA Style

Zhao J, Dong S, Fang Z. A Mixed Finite Element Method for the Multi-Term Time-Fractional Reaction–Diffusion Equations. Fractal and Fractional. 2024; 8(1):51. https://doi.org/10.3390/fractalfract8010051

Chicago/Turabian Style

Zhao, Jie, Shubin Dong, and Zhichao Fang. 2024. "A Mixed Finite Element Method for the Multi-Term Time-Fractional Reaction–Diffusion Equations" Fractal and Fractional 8, no. 1: 51. https://doi.org/10.3390/fractalfract8010051

APA Style

Zhao, J., Dong, S., & Fang, Z. (2024). A Mixed Finite Element Method for the Multi-Term Time-Fractional Reaction–Diffusion Equations. Fractal and Fractional, 8(1), 51. https://doi.org/10.3390/fractalfract8010051

Article Metrics

Back to TopTop