Approximation of Caputo time-fractional diffusion equation using redefined cubic exponential B-spline collocation technique

Mohammad Tamsir; Neeraj Dhiman; Deependra Nigam; Anand Chauhan; Mohammad Tamsir; Neeraj Dhiman; Deependra Nigam; Anand Chauhan

doi:10.3934/math.2021226

AIMS Mathematics

2021, Volume 6, Issue 4: 3805-3820. doi: 10.3934/math.2021226

Previous Article Next Article

Research article

Approximation of Caputo time-fractional diffusion equation using redefined cubic exponential B-spline collocation technique

1.
Department of Mathematics, College of Science, Jazan University, Jazan, Saudi Arabia
2.
Department of Mathematics, Graphic Era Hill University, Dehradun, India
3.
Department of Mathematics, Graphic Era University, Dehradun, India
4.
D. A. V. (P. G.) College, Dehradun, India

Received: 04 November 2020 Accepted: 10 January 2021 Published: 28 January 2021
MSC : 35XX, 65N12

The purpose of this work is to find the numerical solution of the Caputo time-fractional diffusion equation using the modified cubic exponential B-spline (CExpB-spline) collocation technique. First, the CExpB-spline functions are modified and then used to discretize the space derivatives. Three numerical examples are considered for checking the efficiency and accuracy of the method. The obtained results are compared with those reported earlier showing that the present technique gives highly accurate results. Von Neumann stability is carried out which gives the guarantee that the technique is unconditionally stable. The rate of convergence is also obtained. Furthermore, this technique is efficient and requires less storage.

Keywords:

Caputo time-fractional diffusion equation,
collocation technique,
CExpB-spline functions,
stability analysis

Citation: Mohammad Tamsir, Neeraj Dhiman, Deependra Nigam, Anand Chauhan. Approximation of Caputo time-fractional diffusion equation using redefined cubic exponential B-spline collocation technique[J]. AIMS Mathematics, 2021, 6(4): 3805-3820. doi: 10.3934/math.2021226

Related Papers:

[1]	Rabia Noureen, Muhammad Nawaz Naeem, Dumitru Baleanu, Pshtiwan Othman Mohammed, Musawa Yahya Almusawa . Application of trigonometric B-spline functions for solving Caputo time fractional gas dynamics equation. AIMS Mathematics, 2023, 8(11): 25343-25370. doi: 10.3934/math.20231293
[2]	Hasim Khan, Mohammad Tamsir, Manoj Singh, Ahmed Hussein Msmali, Mutum Zico Meetei . Numerical approximation of the time-fractional regularized long-wave equation emerging in ion acoustic waves in plasma. AIMS Mathematics, 2025, 10(3): 5651-5670. doi: 10.3934/math.2025261
[3]	Azhar Iqbal, Abdullah M. Alsharif, Sahar Albosaily . Numerical study of non-linear waves for one-dimensional planar, cylindrical and spherical flow using B-spline finite element method. AIMS Mathematics, 2022, 7(8): 15417-15435. doi: 10.3934/math.2022844
[4]	Zunyuan Hu, Can Li, Shimin Guo . Fast finite difference/Legendre spectral collocation approximations for a tempered time-fractional diffusion equation. AIMS Mathematics, 2024, 9(12): 34647-34673. doi: 10.3934/math.20241650
[5]	M. S. Hashmi, Rabia Shikrani, Farwa Nawaz, Ghulam Mustafa, Kottakkaran Sooppy Nisar, Velusamy Vijayakumar . An effective approach based on Hybrid B-spline to solve Riesz space fractional partial differential equations. AIMS Mathematics, 2022, 7(6): 10344-10363. doi: 10.3934/math.2022576
[6]	Abdul-Majeed Ayebire, Inderpreet Kaur, Dereje Alemu Alemar, Mukhdeep Singh Manshahia, Shelly Arora . A robust technique of cubic Hermite splines to study the non-linear reaction-diffusion equation with variable coefficients. AIMS Mathematics, 2024, 9(4): 8192-8213. doi: 10.3934/math.2024398
[7]	Shafeeq Rahman Thottoli, Mohammad Tamsir, Mutum Zico Meetei, Ahmed H. Msmali . Numerical investigation of nonlinear extended Fisher-Kolmogorov equation via quintic trigonometric B-spline collocation technique. AIMS Mathematics, 2024, 9(7): 17339-17358. doi: 10.3934/math.2024843
[8]	Seherish Naz Khalid Ali Khan, Md Yushalify Misro . Hybrid B-spline collocation method with particle swarm optimization for solving linear differential problems. AIMS Mathematics, 2025, 10(3): 5399-5420. doi: 10.3934/math.2025249
[9]	Emre Kırlı . A novel B-spline collocation method for Hyperbolic Telegraph equation. AIMS Mathematics, 2023, 8(5): 11015-11036. doi: 10.3934/math.2023558
[10]	Shanshan Wang . Split-step quintic B-spline collocation methods for nonlinear Schrödinger equations. AIMS Mathematics, 2023, 8(8): 19794-19815. doi: 10.3934/math.20231009

Abstract

1. Introduction

Fractional differential equations have been applied in modelling of many fields of physics and processes such as earthquakes, optics, finance, hydrology, traffic flow, fluid mechanics, fractional kinetics, mathematical biology, measurement of visco-elastic material properties, electrical network, electro-chemistry electro-magnetic, signal processing, control theory, acoustics, material sciences ^{[2,5,8,9,17,22,39,41,42,43,44,45,46]}. The nonlocal property is the most important advantage of these equations showing state of a complex system does not depend only on its current state but also depends on its all-previous states. Due to this reason, the fractional calculus has becoming more and more popular.

The fractional diffusion equations (FDEs) are generalized forms of classical diffusion equations. These equations have been solved by many researchers. For example, authors of ^{[6,11,24,26,32]} have been applied finite element method, compact FDM, Crank-Nicolson FDM, B-spline based method, implicit difference approximation, respectively, to solve FDEs. Khader ^[14] considered an efficient method based upon Chebyshev approximations. Murio ^[20] developed an implicit unconditionally stable method. Sun et al. ^[25] applied a semi-analytical FEM to solve a class of these equations. Tadjeran et al. ^[27] applied aforesaid method with spatial extrapolation to solve a class of variable coefficients FDEs. Lin and Xu ^[15] constructed a method based on FDM and Legendre spectral method while Ghanbari and Atangana ^[7] presented a method based on the product-integration rule for solving aforesaid equations. Çelik and Duman ^[3] examined a Crank-Nicolson method with the Riesz fractional derivative. Zhai and Feng ^[35] introduced a block-centered FDM for solving these equations.

Additionally, Zhai et al. ^[34,36] constructed unconditionally stable FDMs to solve the time-space fractional diffusion and three-dimensional time-fractional subdiffusion equations, respectively. Wu and Zhai ^[33] proposed a high order FDM to solve the 2D time-fractional convection-dominated diffusion equation while Zhai et al. ^[37,38] proposed a high-order compact FDM and ADI method to solve the 3D fractional convection-diffusion equation. Furthermore, Hanert ^[10] proposed a flexible numerical method to discretize the space-time FDEs, where pseudo-spectral expansion has been used for discretization of the time derivative and either high order pseudo-spectral or low order FE and FD have been used for space derivative. Murillo and Yuste ^[19] considered explicit difference scheme, where three-point centered formula has been used to approximate the spatial derivative. Verma et al. ^[31] studied nonlinear diffusion equations analytically and numerically using the classical Lie symmetry method.

In this paper, we consider a Caputo time-fractional diffusion equation

$\begin{eqnarray} {}^{c} D_{t}^{\alpha} u\left(x, t_n \right) = \frac{{\partial }^2u\left(x, t\right)}{\partial x^2}+f\left(x, t\right), \quad 0\le x\le 1, \ 0\le t\le T, \end{eqnarray}$

(1.1)

subject to the initial condition

$\begin{eqnarray} u\left(x, 0\right) = \varphi(x), \quad 0\le x \le 1, \end{eqnarray}$

(1.2)

the Dirichlet boundary conditions

$\begin{eqnarray} u\left(0, t\right) = 0, \ \ u\left(1, t\right) = 0, \quad 0\le t \le T, \end{eqnarray}$

(1.3)

where ${}^{c} D_{t}^{\alpha} u\left(x, t_n \right)$ denotes the Caputo time-fractional derivative ^[40] as

$\begin{eqnarray} {}^{c} D_{t}^{\alpha} u\left(x, t_n \right) = \frac{1}{\Gamma \left(1-\alpha \right)} \int^t_0{{\left(t-\zeta \right)}^{-\alpha }\frac{\partial u\left(x, \zeta \right)}{\partial \zeta }}d\zeta. \end{eqnarray}$

(1.4)

2. Discretization of the problem

First, the Caputo time-fractional derivative is discretized. For this, we partition the time domain $[0, T]$ uniformly as $0 = t_0 < t_1, ..., < t_N = T$ , where interval $\Delta t = t_n-t_{n-1} = \frac{T}{N}$ for $n = 1, 2, \dots, N$ . Here N is the number of time mesh. By the definition of Caputo derivative at time $t = t_n$ , the time-fractional derivative is approximated as

$\begin{eqnarray} {}^{c} D_{t}^{\alpha} u\left(x, t_n \right) = \frac{1}{\Gamma \left(1-\alpha \right)} \int^{t_n}_0{{\left(t_n-\zeta \right)}^{-\alpha }\frac{\partial u\left(x, \zeta \right)}{\partial \zeta }}d\zeta\\ = \frac{\Delta t^{-\alpha }}{\mathrm{\Gamma }\left(2-\alpha \right)}\sum^{n-1}_{l = 0}{\left(u\left(x, t_{n-l}\right)-u\left(x, t_{n-l-1}\right)\right)\bar{b}_l}+{T_r}^n_{\Delta t}, \end{eqnarray}$

(2.1)

where $\bar{b}_l = {\left(l+1\right)}^{1-\alpha }-l^{1-\alpha }\ \forall \ 0\le l \le N$ and ${T_r}^n_{\Delta t}$ gives the truncation error $\le \bar{c}_u\Delta t^{2-\alpha }$ , where the constant $\bar{c}_u$ is only related to $u(x, t)$ .

Lemma 3.1 The below relations should be hold for the coefficients $\bar{b}_l$

$(1)$ $\bar{b}_0 = 1$ ,

$(2)$ $\bar{b}_l > 0, \quad \forall \ \ 0\le l\le N,$

$(3)$ $\bar{b}_{l-1} > \bar{b}_l, \ \forall \ 1\le l\le N$ .

Proof. For the proof of Lemma 3.1, see ^[16].

Next, we use CExpB-spline collocation technique to discretize the space derivatives. First, we partition the space domain $[a, b]$ uniformly as $a = x_0 < x_1, ..., x_M = b$ with space size $\Delta x = h = x_{m+1}-x_m = \frac{b-a}{M}\,$ where $m = 0, 1, \dots, M$ . For discrete form, we denote $u_m^n = u(x_m, t_n)$ for $m = 0, 1, ..., M$ and $n = 0, 1, ..., N$ . The CExpB-spline functions ${Ep}_m\left(x\right)$ for $m = -1, \ 0, \ 1, \dots, M+1$ are defined as ^[28,29,30]:

$\begin{eqnarray} Ep_m \left( x \right) = \frac{1}{{h^3 }}\left\{ {\begin{array}{*{20}c} {\hat \gamma _5 \left( {\left( {x_{m - 2} - x} \right) - \frac{1}{{\hat \kappa }}\left( {\sinh \left( {\hat \kappa \left( {x_{m - 2} - x} \right)} \right)} \right)} \right), \quad \quad \ x \in \left[ {x_{m - 2} , x_{m - 1} } \right)} \\ {\hat \gamma _1 + \hat \gamma _2 \left( {x_m - x} \right) + \hat \gamma _3 {\mathop{\rm e}\nolimits} ^{\hat \kappa \left( {x_m - x} \right)} + \hat \gamma _4 {\mathop{\rm e}\nolimits} ^{ - \hat \kappa \left( {x_m - x} \right)} , \ x \in \left[ {x_{m - 1} , x_m } \right)} \\ {\hat \gamma _1 + \hat \gamma _2 \left( {x - x_m } \right) + \hat \gamma _3 {\mathop{\rm e}\nolimits} ^{ - \hat \kappa \left( {x_m - x} \right)} + \hat \gamma _4 {\mathop{\rm e}\nolimits} ^{\hat \kappa \left( {x_m - x} \right)} , \ x \in \left[ {x_m , x_{m + 1} } \right)\, } \\ {\hat \gamma _5 \left( {\left( {x - x_{m + 2} } \right) - \frac{1}{{\hat \kappa }}\left( {\sinh \left( {\hat \kappa \left( {x - x_{m + 2} } \right)} \right)} \right)} \right), \quad \quad x \in \left[ {x_{m + 1} , x_{m + 2} } \right)} \\ {0, \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad otherwise} \\ \end{array}} \right. \end{eqnarray}$

(2.2)

where

$\begin{eqnarray} \begin{array}{l} \hat \gamma _1 = \frac{{\hat \kappa hc}}{{\hat \kappa hc - s}}, \ \hat \gamma _2 = \frac{{\hat \kappa }}{2}\left( {\frac{{c\left( {c - 1} \right) + s^2 }}{{\left( {\hat \kappa hc - s} \right)\left( {1 - c} \right)}}} \right), \ \hat \gamma _3 = \frac{1}{4}\left( {\frac{{{\mathop{\rm \left( {1 - c} \right) e}\nolimits} ^{ - \hat \kappa h} + s\left( {{\mathop{\rm e}\nolimits} ^{ - \hat \kappa h} - 1} \right)}}{{\left( {\hat \kappa hc - s} \right)\left( {1 - c} \right)}}} \right), \\ \hat \gamma _4 = \frac{1}{4}\left( {\frac{{{\mathop{\rm e}\nolimits} ^{\hat \kappa h} \left( {c - 1} \right) + s\left( {{\mathop{\rm e}\nolimits} ^{\hat \kappa h} - 1} \right)}}{{\left( {\hat \kappa hc - s} \right)\left( {1 - c} \right)}}} \right), \ \hat \gamma _5 = \frac{{\hat \kappa }}{{2\left( {\hat \kappa hc - s} \right)}}, \ s = \sinh (\hat \kappa h), \ c = \cosh (\hat \kappa h). \\ \end{array} \end{eqnarray}$

In Eq (2.2), the free parameter $\hat \kappa$ have been used for obtaining the different forms of CExpB-spline functions. The set of ${Ep}_m\left(x\right)\ \forall \ m = -1, \ 0, \ 1, \dots, M+1$ forms a basis over the problem domain. We assume that the approximation $u^M$ to the exact $u(x, t)$ at the point $(x_m, t_n)$ is expressed in terms of linear combinations of the CExpB-spline functions and unknown time-dependent quantities as follows:

$\begin{equation} u^M\left(x_m, t_n\right) = \sum\limits^{M+1}_{i = -1} C_i(t) Ep_i(x), \end{equation}$

(2.3)

where $C_i\left(t\right)$ are the unknown quantities which we have to evaluate for the approximated solution $u^M\left(x, t\right)$ at $\left(x_i, t_j\right)$ . Since each CExpB-spline covers four elements, so each element is covered by four CExpB-splines. So the variation of the $u^M(x, t)$ , over the element, can be written as:

$\begin{equation} u^M\left(x, t\right) = \sum\limits^{m+1}_{i = m-1} C_i(t) Ep_i(x). \end{equation}$

(2.4)

Using Eq (2.4), the $u(x, t)$ and its first two derivatives at the knots in terms of $C_m^n$ are given as:

$\begin{eqnarray} u^n_m = \frac{{s - \hat \kappa h}}{{2\left( {\hat \kappa hc - s} \right)}} C_{m-1}^n+ C_m^n + \frac{{s - \hat \kappa h}}{{2\left( {\hat \kappa hc - s} \right)}} C_{m+1}^n, \end{eqnarray}$

(2.5)

$\begin{eqnarray} {\left(u_x\right)}^n_m = -\frac{{\hat \kappa (c - 1)}}{{2\left( {\hat \kappa hc - s} \right)}} \left(C_{m-1}^n- C_{m+1}^n\right), \end{eqnarray}$

(2.6)

$\begin{eqnarray} {\left(u_{xx}\right)}^n_m = \frac{{\hat \kappa ^2 s}}{{2\left( {\hat \kappa hc - s} \right)}} \left( C_{m-1}^n-2 C_m^n + C_{m+1}^n\right), \end{eqnarray}$

(2.7)

Now, we modify the CExpB-spline basis functions which generate a new set of CExpB-spline basis functions. The procedure, for modifying the basis functions, is given as ^[1,12,13]:

$\begin{eqnarray} \left\{ \begin{array}{lr} {\hat{Ep}_0(x) = 4{Ep}_{-1}(x)+{Ep}_0(x)}, \\ {\hat{Ep}_1(x) = {Ep}_1(x)-{Ep}_{-1}(x) }, \\ {\hat{Ep}_m(x) = {Ep}_m(x), \ m = 2, 3, \dots , M-2}, \\ {\hat{Ep}_{M-1}(x) = {Ep}_{M-1}(x)-{Ep}_{M+1}(x)}, \\ {\hat{Ep}_M(x) = {Ep}_M(x)+4{Ep}_{M+1}(x)}, \end{array} \right. \end{eqnarray}$

(2.8)

where $\{\hat{Ep}_0(x), \hat{Ep}_1(x)\dots \dots \hat{Ep}_M(x)\}$ forms a basis over the problem domain. Next, the modified form of the approximated solutions, as a linear combination of modified CExpB-spline functions, is given by

$\begin{eqnarray} \left\{ \begin{array}{lr} {u^M(x_0, t_n) = 0}, \ for \ m = 0, \\ {u^M(x_m, t_n) = \sum^M_{j = 0}{\hat{Ep}_j(x) C_j(t)}, } \ for \ m = 1, 2, \dots , M-1, \\ {u^M(x_M, t_n) = 0}, \ for \ m = M. \end{array} \right. \end{eqnarray}$

(2.9)

Next, the initial vector $\left(C^0_m, \ m = 0, \ 1, \dots, \ M\right)$ can be obtained by initial condition and boundary values of its derivatives as:

$\begin{eqnarray} \left\{ \begin{array}{lr} {u^M(x_0, 0) = 0}, \ for \ m = 0, \\ {u^M(x_m, 0) = {\varphi }_m}, \ for \ m = 1, 2, \dots , M-1, \\ {u^M(x_M, 0) = 0}, \ for \ m = M. \end{array} \right. \end{eqnarray}$

(2.10)

Eq (2.10) yields a $(M+1) \times (M+1)$ system as:

$\begin{equation} \left[ {\begin{array}{*{20}c} \frac{\hat \kappa h (c-1)} {{\left( {\hat \kappa hc - s} \right)}} \ & 0 & & & & \\ \frac{{s - \hat \kappa h}}{{2\left( {\hat \kappa hc - s} \right)}} & \ 1 \ & \ \frac{{s - \hat \kappa h}}{{2\left( {\hat \kappa hc - s} \right)}} & & & \\ 0 & \ \frac{{s - \hat \kappa h}}{{2\left( {\hat \kappa hc - s} \right)}} & \ 1 & \ \ \frac{{s - \hat \kappa h}}{{2\left( {\hat \kappa hc - s} \right)}} & & \\ & & \ddots & \ddots & \ddots & \\ & & & \frac{{s - \hat \kappa h}}{{2\left( {\hat \kappa hc - s} \right)}} & \ 1 & \ \frac{{s - \hat \kappa h}}{{2\left( {\hat \kappa hc - s} \right)}} \\ & & & & \ 0 & \ \frac{\hat \kappa h (c-1)} {{\left( {\hat \kappa hc - s} \right)}} \\ \end{array}} \right] \left[ {\begin{array}{*{20}c} {C_1^0 } \\ {C_2^0 } \\ {C_3^0 } \\ \vdots \\ {C_{M - 1}^0 } \\ {C_M^0 } \\ \end{array}} \right] = \left[ {\begin{array}{*{20}c} 0 \\ {\varphi _1 } \\ {\varphi _2 } \\ \vdots \\ {\varphi _{M - 1} } \\ 0 \\ \end{array}} \right]. \end{equation}$

(2.11)

Now, discretizing the Eq (1.1) by using (2.1) and Crank-Nicolson method, we have

$\frac{\Delta t^{-\alpha }}{\mathrm{\Gamma }\left(2-\alpha \right)}\sum\limits_{l = 0}^{n-1} \left(u_m^{n-l}-u_m^{n-l-1}\right) (\bar{b}_l) = \frac{1}{2}\left\{\left(u_{x x}\right)_{m}^{n}+\left(u_{x x}\right)_{m}^{n-1}\right\} + \frac{1}{2} \left(f_m^n+f_m^{n-1}\right), \\ m = 0, 1, ..., M, \ n = 1, 2, ..., N.$

(2.12)

Using Eqs (2.4) and (2.6) in Eq (2.12) and simplifying the terms, we have

$\begin{eqnarray} \left( {\frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha } \right)}}k_0 \omega_1 - 0.5\omega_2 } \right)C_{m - 1}^n + \left( {\frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha } \right)}}k_0 +\omega_2} \right)C_m^n + \left( {\frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha} \right)}}k_0 \omega_1} \right. \\ \left. {-0.5\omega_2 } \right)C_{m + 1}^n = \frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha } \right)}}\left[ {\sum\limits_{l = 1}^{n - 1} {\left( {\omega_1 C_{m - 1}^{l - 1} +C_m^{l - 1} + \omega_1 C_{m + 1}^{l - 1} } \right)} } \right.\left( {\bar b_{l - 1} - \bar b_l } \right) + \\ \left. {\bar b_{n - 1} \left( {\omega_1 C_{m - 1}^0 } \right. + \left. {C_m^0 + \omega_1 C_{m + 1}^0 } \right)} \right] + 0.5 \omega_2 \left( {C_{m - 1}^{n - 1}- 2 C_m^{n - 1} + C_{m + 1}^{n - 1} } \right) + \\ \frac{1}{2}\left( {f_m^n + f_m^{n - 1} } \right), \ m = 1, 2, ..., M, \ n = 1, 2, ..., N. \end{eqnarray}$

(2.13)

where

$\begin{eqnarray} \begin{array}{l} \omega _1 = \frac{{s - \hat \kappa h}}{{2\left( {\hat \kappa hc - s} \right)}}, \ \text{and} \ \omega_2 = \frac{{\hat \kappa ^2 s}}{{2\left( {\hat \kappa hc - s} \right)}}. \end{array} \end{eqnarray}$

For $m = 0$ , from Eq (2.13), we get

$\begin{eqnarray} \left( {\frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha } \right)}}\left(1+2\omega_1 \right) \bar b_0} \right)C_0^n = \left(1+2\omega_1 \right) \frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha } \right)}}\left[ {\sum\limits_{l = 1}^{n - 1} {\left( {\left( {\bar b_{l - 1} - \bar b_l } \right) } \right.} C_0^{l - 1} + \left. {\bar b_{n - 1} C_0^0 } \right)} \right] \\ + \frac{1}{2}\left( {f_0^n + f_0^{n - 1} } \right), \ \ n = 1, 2, ..., N. \end{eqnarray}$

(2.14)

For m = M, from Eq (2.13), we get

$\begin{eqnarray} \left( {\frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha } \right)}}\left(1+2\omega_1 \right) \bar b_0} \right)C_M^n = \left(1+2\omega_1 \right) \frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha } \right)}}\left[ {\sum\limits_{l = 1}^{n - 1} {\left( {\left( {\bar b_{l - 1} - \bar b_l } \right) } \right.} C_M^{l - 1} + \left. {\bar b_{n - 1} C_M^0 } \right)} \right] \\ + \frac{1}{2}\left( {f_M^n + f_M^{n - 1} } \right), \ \ n = 1, 2, ..., N. \end{eqnarray}$

(2.15)

At time step $t_{n}, \ n = 1, 2, ..., N$ , the Eqs (2.13)–(2.15) can be formulated into a $(M+1)\times(M+1)$ linear system as follows:

$\begin{equation} \left( {\begin{array}{*{20}c} \hat A & {0} & {0} & {} & {} & {} \\ {\hat B_1 } & {\hat B_2 } & {\hat B_1 } & {} & {} & {} \\ {} & {\hat B_1 } & {\hat B_2 } & {\hat B_1 } & {} & {} \\ {} & {} & \ddots & \ddots & \ddots & {} \\ {} & {} & {} & {\hat B_1 } & {\hat B_2 } & {\hat B_1 } \\ {} & {} & {} & {0} & {0} & \hat A \\ \end{array}} \right) \left( {\begin{array}{*{20}c} {C_0^n } \\ \begin{array}{l} C_1^n \\ C_2^n \\ \end{array} \\ \vdots \\ {C_{M - 1}^n } \\ {C_M^n } \\ \end{array}} \right) = \left( {\begin{array}{*{20}c} {R_0 } \\ {R_1 } \\ {R_2 } \\ \vdots \\ {R_{M - 1} } \\ {R_M } \\ \end{array}} \right), \end{equation}$

(2.16)

where

$\begin{eqnarray} \begin{array}{l} \hat A = {\frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha } \right)}}\left(1+2\omega_1 \right) \bar b_0}, \ \hat B_1 = \frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha } \right)}}k_0 \omega_1 - 0.5\omega_2 , \ \hat B_2 = \frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha } \right)}}k_0 +\omega_2 , \\ R_0 = \frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha } \right)}}\left[ {\sum\limits_{l = 1}^{n - 1} {\left( {\left( {\bar b_{l - 1} - \bar b_l } \right)\left( {1+2\omega_1} \right)C_0^{l - 1} + } \right.} } \right.\left. {\left. {\bar b_{n - 1} \left( {1+2\omega _1} \right)C_0^0 } \right)} \right]+ \frac{1}{2}\left( {f_0^n + f_0^{n - 1} } \right), \\ n = 1, 2, ..., N, \\ R_m = \frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha } \right)}}\left[ {\sum\limits_{l = 1}^{n - 1} {\left( {\left( {\omega_1 C_{m - 1}^{l - 1} + C_m^{l - 1} + \omega_1 C_{m + 1}^{l - 1} } \right)} \right.} } \right.\left( {\bar b_{l - 1} - \bar b_l } \right) + \bar b_{n - 1} \left( {\omega_1 C_{m - 1}^0 + C_m^0 + } \right. \\ \left. {\left. {\left. {\omega_1 C_{m + 1}^0 } \right)} \right)} \right] + 0.5\left( {\omega_2 C_{m - 1}^{n - 1} -2 \omega _2 C_m^{n - 1} + \omega _2 C_{m + 1}^{n - 1} } \right) + \frac{1}{2}\left( {f_m^n + f_m^{n - 1} } \right), \\ m = 1, 2, ..., M - 1, n = 1, 2, ..., N, \\ R_M = \frac{{\Delta t^{ - \alpha } }}{{\Gamma \left( {2 - \alpha } \right)}}\left[ {\sum\limits_{l = 1}^{n - 1} {\left( {\left( {\bar b_{l - 1} - \bar b_l } \right)\left( {1+2\omega_1} \right)C_M^{l - 1} + } \right.} } \right.\left. {\left. {\bar b_{n - 1} \left( {1+2\omega_1 } \right)C_M^0 } \right)} \right] + \frac{1}{2}\left( {f_M^n + f_M^{n - 1} } \right), \\ n = 1, 2, ..., N.\\ \end{array} \end{eqnarray}$

3. Stability analysis

The von-Neumann stability analysis ^[4,18,21,23] have been done in this section. Taking $f\left(x, t\right) = 0$ in Eq (1.1) and discretizing, we get

(3.1)

Simplifying and rearranging the terms, we get

$\begin{eqnarray} \left(\frac{\Delta t^{-\alpha}}{\Gamma\left(2-\alpha \right)} \bar{b}_0 \omega_1-0.5 \omega _2\right) C_{m-1}^n+\left(\frac{\Delta t^{-\alpha}}{\Gamma\left(2-\alpha \right)} \bar{b}_0 + \omega_2 \right) C_m^n+ \left({\frac{\Delta t^{-\alpha}}{\Gamma\left(2-\alpha \right)} \bar{b}_0 \omega_1} \right. \\ - \left. 0.5 \omega_2\right) C_{m+1}^n = 0.5 \omega_2 \left(C_{m-1}^{n-1}- 2C_m^{n-1} + C_{m+1}^{n-l}\right) + \frac{\Delta t^{-\alpha }}{\Gamma \left(2-\alpha \right)} \sum\limits_{l = 1}^{n-1} \Big(\omega_1 C_{m-1}^{n-l}\\ + C_m^{n-l} +\omega_1 C_{m+1}^{n-l}\Big) (\bar{b}_{l-1} -\bar{b}_l) \left. +\frac{\Delta t^{-\alpha }}{\Gamma \left(2-\alpha \right)} \bar{b}_{n-1} \left(\omega_1 C_{m-1}^0+ C_m^0+ \omega_1 C_{m+1}^0 \right), \right. \end{eqnarray}$

(3.2)

which can be written as

$\begin{eqnarray} A C_{m-1}^n+ B C_m^n+ A C_{m+1}^n = D C_{m-1}^{n-1}+E C_m^{n-1}+ D C_{m+1}^{n-1}+ F, \end{eqnarray}$

(3.3)

where

$\begin{eqnarray} A = \left(\frac{\Delta t^{-\alpha }}{\Gamma \left(2-\alpha \right)} \bar{b}_0 \omega_1 - 0.5 \omega_2\right), \ B = \left(\frac{\Delta t^{-\alpha }}{\mathrm{\Gamma }\left(2-\alpha \right)} \bar{b}_0+ \omega_2\right), \ D = 0.5 \omega_2, \ S = -\omega_2 \end{eqnarray}$

and

$\begin{eqnarray} F = \frac{\Delta t^{-\alpha }}{\Gamma \left(2-\alpha \right)} \sum\limits_{l = 1}^{n-1} \Big(\omega_1 C_{m-1}^{n-l} + C_m^{n-l} +\omega_1 C_{m+1}^{n-l}\Big) (\bar{b}_{l-1} -\bar{b}_l)\\ + \frac{\Delta t^{-\alpha }}{\Gamma \left(2-\alpha \right)} \bar{b}_{n-1} \left(\omega_1 C_{m-1}^0+ C_m^0+ \omega_1 C_{m+1}^0 \right). \end{eqnarray}$

Now we consider one Fourier mode out of the full solution $C_m^n = \delta ^n e^{ki \phi}$ as trial solutions at a given point $x_i$ , where $\phi = \theta h$ . The $\theta$ and $h$ are the mode number and element size respectively, and $k = \sqrt{-1}$ . Substituting the trial solution in above equation and simplifying the terms, we get

$\begin{eqnarray} \delta ^n \left( {2A\cos \phi + B} \right) = \delta ^{n - 1} \left( {2D\cos \phi + D} \right) + \frac{{\Delta t^{ - \alpha } }}{{\Gamma (2 - \alpha)}}\left( {1+2\omega_1 \cos \phi} \right) \bar F, \label{eqsa7} \end{eqnarray}$

where $\bar F = \sum\limits_{l = 1}^{n - 1} {\left({\bar b_{l-1} - \bar b_l } \right)} \delta ^{n - l} + \bar b_{n - 1} \delta ^0.$

Now we define

$\begin{eqnarray} \delta _{\max }^{n - 1} = \mathop {\max }\limits_{0 \le j \le n - 1} |\delta ^j| \end{eqnarray}$

(3.4)

Using (3.4), we get

$\begin{eqnarray} \bar F = \sum\limits_{l = 1}^{n - 1} {\left( {\bar b_{l-1} - \bar b_l} \right)} \delta ^{n - l} + \bar b_{n - 1} \delta ^0 \le \sum\limits_{l = 1}^{n - 1} {\left( {\bar b_{l-1} - \bar b_l + \bar b_{n - 1} } \right)} \delta _{\max }^{n - 1} = \bar b_0 \delta _{\max }^{n - 1} . \end{eqnarray}$

(3.5)

Now, using above equation and the values of $A, \ B, \ D$ and $E$ in Eq (3.3), we have

$|\delta| \leq\left|\frac{\frac{\Delta t^{-\alpha}}{\Gamma(2-\alpha)}\left(1+2 \omega_{1} \cos \phi\right) \bar{b}_{0}-\omega_{2}\left(1+2 \sin ^{2}(\phi)\right)}{\frac{\Delta t^{-\alpha}}{\Gamma(2-\alpha)}\left(1+2 \omega_{1} \cos \phi\right) \bar{b}_{0}+\omega_{2}\left(1+2 \sin ^{2}(\phi)\right)}\right|.$

(3.6)

From (3.6), we get $|\delta| \le 1$ , and hence the method is unconditionally stable for the discretized system of the Caputo time-fractional diffusion equation.

4. Computational results and discussions

In this section, we consider numerical examples of the Caputo time-fractional diffusion equation in order to check the accuracy and efficiency of the method. We use the following error norms :

$\begin{eqnarray} L_2 = \sqrt {h\sum\limits_{j = 0}^M {\left| {u_j^{exact} - u_j^{numerical} } \right|^2 } } , \, \, \ \ L_{max} = \mathop {\max }\limits_{0 \le j \le M} |u_j^{exact} - u_j^{numerical} |. \end{eqnarray}$

(4.1)

The ROC is calculated by

$\begin{eqnarray} ROC = \frac{\log \left(E^{h_1} / E^{h_2} \right)} {\log \left({h_1} / {h_2} \right)} \ {\text{and} \ ROC = \frac{\log \left(E^{k_1} / E^{k_2} \right)} {\log \left({k_1} / {k_2} \right)}, } \end{eqnarray}$

(4.2)

where, $E^{h_1}$ and $E^{h_2}$ represent the errors at space mesh sizes $h_1$ and $h_2$ , respectively and $E^{k_1}$ and $E^{k_2}$ represent the errors at time mesh sizes $k_1$ and $k_2$ , respectively.

4.1. Example 1

First, we consider the Caputo time-fractional diffusion equation

$\begin{equation} {}^{c} D_{t}^{\alpha} u\left(x, t_n \right) = \frac{{\partial}^2 u\left(x, t\right)}{\partial t^2}+ f \left(x, t\right), \ \ \ \ \ \ x \in \left[0, 1 \right], \ \ t\ge 0, \ \ 0 \lt \alpha \lt 1, \end{equation}$

(4.3)

with the exact solution

$u\left(x, t\right) = t^2 \sin(2\pi x),$

where

$f\left(x, t\right) = \left(\frac{2} {\Gamma \left(3-\alpha \right)} t^{2-\alpha} + 4{\pi }^2t^2\right)\sin\left(2\pi x\right).$

We solve the Caputo time-fractional diffusion Eq (4.3) for different values of $M$ and $N$ . Table 1 shows the comparison of the present method with the finite element method ^[6] and cubic B-spline collocation method ^[24] for different values of $M$ , fixed $\Delta t = 0.001$ at the time-fractional order $\alpha = 0.5\$ in terms of $L_2$ and $L_{max}$ error norms. This table shows that the present method gives better results than the results obtained by those available in ^[6,24]. Moreover, one can notice from that the error norms are decreasing as we increase the space as well as time mesh sizes. The analytical and approximate solutions together with the absolute errors are presented in for $\hat \kappa = 11$ , $\Delta x = 0.001$ , and $\alpha = 0.5$ at $\Delta t = 0.2$ , $\Delta t = 0.1$ , and $\Delta t = 0.05$ , respectively. shows the comparison between analytical and approximate solutions for $M = 64$ , $\Delta t = 0.001$ , and $\alpha = 0.5$ at different $t$ . From these figures, we notice that there is an excellent agreement between analytical and approximate solutions. Also, the absolute errors are decreasing on increasing the time interval.

Table 1. Comparison of present results and the results available in ^[6,24], for Example 1.

$M$	Present Method $L_{max}$	Present Method $L_2$	Cubic B-spline collocation method ^[24] $L_{max}$	Finite element method ^[6] $L_2$
4	1.28e-03	9.06e-03	1.72e-01	1.09e-01
8	2.73e-04	1.93e-04	4.80e-02	3.20e-02
16	1.09e-04	7.67e-05	1.23e-02	8.42e-03
32	8.65e-05	6.12e-05	3.08e-03	2.15e-03
64	8.23e-05	5.82e-05	7.70e-04	5.42e-04

| Show Table

DownLoad: CSV

Figure 1. Analytical and approximate solutions together with absolute errors for

$\hat \kappa = 11$ ,

$\Delta x = 0.001$ , and

$\alpha = 0.5$ at (a)

$\Delta t = 0.2$ , (b)

$\Delta t = 0.1$ , and (c)

$\Delta t = 0.05$ (Example 1).

DownLoad: Full-Size Img PowerPoint

Figure 2. Analytical and approximate

$u(x, t)$ for

$M = 64$ ,

$\Delta t = 0.001$ and

$\alpha = 0.5$ at different

$t$ (Example 1).

DownLoad: Full-Size Img PowerPoint

4.2. Example 2

Next, we consider the Caputo time-fractional diffusion equation

$\begin{equation} {}^{c} D_{t}^{\alpha} u\left(x, t_n \right) = \frac{{\partial }^2u\left(x, t\right)}{\partial x^2}+f\left(x, t\right), \ \ x\in \left[0, 1\right], \ \ t \ge 0, \ \ 0 \lt \alpha \lt 1, \end{equation}$

(4.4)

with the exact solution

$u\left(x, t\right) = x^2{\left(x-1\right)}^2e^xt^{(2+\alpha )},$

where

$f\left(x, t\right) = 0.5 t^2 e^x x^2{(x-1)}^2 \Gamma \left(\alpha +3\right)-t^{(2+\alpha )} e^x (x^4+6x^3+x^2-8x+2).$

We solve the Caputo time-fractional diffusion Eq (4.4) for various values of $M$ . shows the maximum absolute error norms $L_{max}$ for different values of time-fractional orders i. e. $\alpha = 0.2, \ 0.5\$ and $0.8$ at fixed time mesh size of $N = 50$ for different space mesh sizes. One can see from and that the error norms are decreasing as we increase the mesh sizes. As we can notice from this table that the present method is of order $O\left((\Delta x)^2, (\Delta t)^{1+\alpha}\right)$ . The approximate solutions for time levels $t =$ 0.1, 0.4, 0.6 and 0.8 are plotted in for $M = 32, \ N = 50$ and time-fractional order $\alpha = 0.5$ . One can notice that the amplitude of the approximate solutions is increasing on incresing the time $t$ .

Table 2. The

$L_{max}$ error norms and convergence order for different

$M$ and

$\alpha$ at fixed

$\ N = 50$ (Example 2).

$M$	$\alpha =0.2$	ROC	$\ \alpha =0.5$	ROC	$\alpha =0.8$	ROC
2	6.67e-01	-	1.94e-01	-	1.287e-02	-
4	1.715e-01	1.96	5.01e-02	1.95	2.987e-03	2.10
8	4.205-02	2.03	1.23e-02	2.02	6.958e-04	2.10
16	1.040e-02	2.02	3.99e-03	1.99	1.710e-04	2.02
32	2.051e-03	2.05	7.81e-04	1.98	4.297e-05	1.99

| Show Table

DownLoad: CSV

Table 3. The

$L_{max}$ error norms and convergence order for different

$N$ and

$\alpha$ at fixed

$\ M = 50$ (Example 2).

$N$	$\alpha =0.2$	ROC	$\ \alpha =0.5$	ROC	$\alpha =0.8$	ROC
2	9.270e-04	-	5.478e-04	-	3.733e-04	-
4	3.927e-04	1.23	1.987e-04	1.46	1.032e-04	1.85
8	1.618e-04	1.27	7.182e-05	1.47	2.733e-05	1.91
16	6.901e-05	1.22	2.609e-05	1.46	7.793e-05	1.81
32	2.950e-05	1.22	9.472e-06	1.46	2.139e-06	1.86

| Show Table

DownLoad: CSV

Figure 3. Numerical solutions for

$M = 32$ ,

$N = 50$ and

$\alpha = 0.5$ at different

$t$ (Example 2).

DownLoad: Full-Size Img PowerPoint

4.3. Example 3

Now, we consider the Caputo time-fractional diffusion equation

$\begin{equation} {}^{c} D_{t}^{\alpha} u\left(x, t_n \right) = \frac{{\partial }^2u \left(x, t\right)} {\partial x^2}+ f\left(x, t\right), \ \ 0 \lt x, \alpha \lt 1, \ t \gt 0, \end{equation}$

(4.5)

where

$f\left(x, t\right) = t^{\beta }e^x\left(\frac{\mathrm{\Gamma }\left(\beta +1\right)}{\mathrm{\Gamma }\left(\beta +1-\alpha \right)}t^{-\alpha }-1\right),$

subject to the initial and boundary conditions

$\ u\left(x, 0\right) = 0, \ \ \ 0 \lt x \lt 1,$

$u\left(0, t\right) = t^{\beta }, \ \ u\left(1, t\right) = et^{\beta }\ \ \ \forall t\ge 0\ ,$

and with the exact solution

$u\left(x, t\right) = e^xt^{\beta }.$

Now, we solve the Caputo time-fractional diffusion Eq (4.5) for different values of $M$ and $N$ . shows the comparison in term of $L_{max}\$ error norms obtained by the present method and the method based on the product-integration (PI) rule presented in ^[7] for $\Delta x = \frac{1}{101}$ , $\beta = 6$ , different values of $N$ and $\alpha$ at $t = 1$ . As we can see that the present method gives more accurate results than the method presented in ^[7]. Also, we notice that the error norms are decreasing as we increase the time mesh sizes. shows the analytical and approximate solutions together with absolute error norms for $\Delta x = \frac{1}{101}, \ \alpha = 0.85, \ \beta = 5, \ \Delta t = 0.01$ at $T = 0.35$ showing that errors are decreasing on increasing the grid sizes.

Table 4. The

$L_{max}$ error norms for different time-fractional orders with

$\Delta x = \frac{1}{101}$ ,

$\beta = 6$ at

$T = 1$ (Example 3).

$N$	Present Method for $\alpha=0.5$	Method based on PI rule ^[7] for $\alpha=0.5$	Present method for $\alpha=0.7$	Method based on PI rule ^[7] for $\alpha=0.7$	Present method for $\alpha=0.9$	Method based on PI rule ^[7] for $\alpha=0.9$
10	2.34e-03	7.63e-03	8.41e-03	1.07e-02	7.59e-03	1.14e-02
20	6.23e-04	2.03e-03	1.25e-03	2.76e-04	2.71e-03	3.53e-03
40	2.54e-04	5.27e-04	6.18e-04	7.03e-04	6.78e-04	8.89e-04
80	7.22e-05	1.36e-04	1.57e-04	1.78e-04	1.54e-04	2.23e-04
160	3.31e-05	3.55e-05	4.49e-05	4.56e-05	4.35e-05	5.67e-05

| Show Table

DownLoad: CSV

Figure 4. Analytical and approximate solutions together with absolute errors for

$p = 0.01$ ,

$\Delta x = 0.001$ , and

$\alpha = 0.5$ at (a)

$\Delta t = 0.08$ , (b)

$\Delta t = 0.04$ , and (c)

$\Delta t = 0.02$ (Example 3).

DownLoad: Full-Size Img PowerPoint

5. Conclusions

A modified CExpB-spline collocation technique has been presented for solving the Caputo time-fractional diffusion equation. The modified CExpB-spline collocation technique is used to discretize the space derivatives. The three examples of the Caputo time-fractional diffusion equation have been considered. The obtained results show that the present method gives more accurate results than the results obtained in ^[6,7,24]. It is observed numerically that the method is second-order accurate in space and $(1+\alpha)$ order in time. The stability analysis shows that the method is unconditionally stable. Moreover, the implementation of the present method is easy and needs low memory storage which is the advantage. The present method can easily be extended for solving higher dimensional fractional PDEs.

Conflict of interest

The authors have no conflict of interest.

References

[1]	A. S. Alshomrani, S. Pandit, A. K. Alzahrani, M. S. Alghamdi, R. Jiwari, A numerical algorithm based on modified cubic trigonometric B-spline functions for computational modelling of hyperbolic-type wave equations, Eng. Computation., 34 (2017), 1257–1276. doi: 10.1108/EC-05-2016-0179
[2]	P. J. Torvik, R. L. Bagley, On the appearance of the fractional derivative in the behavior of real materials, J. Appl. Mech., 51 (1984), 294–298. doi: 10.1115/1.3167615
[3]	C. Çelik, M. Duman, Crank-Nicolson method for the fractional diffusion equation with the Riesz fractional derivative, J. Comput. Phys., 231 (2012), 1743–1750. doi: 10.1016/j.jcp.2011.11.008
[4]	P. K. Chattaraj, S. R. Koneru, B. M. Deb, Stability analysis of finite difference schemes for quantum mechanical equations of motion, J. Comput. Phys., 72 (1987), 504–512. doi: 10.1016/0021-9991(87)90098-2
[5]	W. H. Deng, C. P. Li, The evolution of chaotic dynamics for fractional unified system, Phys. Lett. A, 372 (2008), 401–407. doi: 10.1016/j.physleta.2007.07.049
[6]	N. J. Ford, J. Xiao, Y. Yan, A finite element method for time fractional partial differential equations, Fract. Calc. Appl. Anal., 14 (2011), 454–474.
[7]	B. Ghanbari, A. Atangana, An efficient numerical approach for fractional diffusion partial differential equations, Alex. Eng. J., 59 (2020), 2171–2180. doi: 10.1016/j.aej.2020.01.042
[8]	M. Giona, S. Cerbelli, H. E. Roman, Fractional diffusion equation and relaxation in complex viscoelastic materials, Physica A., 191 (1992), 449–453. doi: 10.1016/0378-4371(92)90566-9
[9]	E. Goldfain, Fractional dynamics, Cantorian space-time and the gauge hierarchy problem, Chaos, Soliton. Fract., 22 (2004), 513–520. doi: 10.1016/j.chaos.2004.02.043
[10]	E. Hanert, On the numerical solution of space-time fractional diffusion models, Comput. Fluids, 46 (2011), 33–39. doi: 10.1016/j.compfluid.2010.08.010
[11]	B. A. Jacobs, High-Order Compact Finite Difference and Laplace Transform Method for the solution of Time-Fractional Heat Equations with Dirchlet and Neumann Boundary conditions, Numer. Meth. Part. D. E., 32 (2016), 1184–1199. doi: 10.1002/num.22046
[12]	R. Jiwari, A. S. Alshomrani, A new algorithm based on modified trigonometric cubic B-splines functions for nonlinear Burgers'-type equations, Int. J. Numer. Method. H., 27 (2017), 1638–1661. doi: 10.1108/HFF-05-2016-0191
[13]	R. Jiwari, S. Pandit, M. E. Koksal, A class of numerical algorithms based on cubic trigonometric B-spline functions for numerical simulation of nonlinear parabolic problems, Comput. Appl. Math., 38 (2019).
[14]	M. M. Khader, On the numerical solutions for the fractional diffusion equation, Communications in Nonlinear Science and Numerical Simulation, Commun. Nonlinear Sci., 16 (2011), 2535–2542. doi: 10.1016/j.cnsns.2010.09.007
[15]	Y. Lin, C. Xu, Finite difference/spectral approximations for the time-fractional diffusion equation, J. Comput. Phys., 225 (2007), 1533–1552. doi: 10.1016/j.jcp.2007.02.001
[16]	F. Liu, P. Zhuang, V. Anh, I. Turner, K. Burrage, Stability and convergence of the difference methods for the space-time fractional advection-diffusion equation, Appl. Math. Comput., 191 (2007), 12–20.
[17]	F. Mainardi, Fractional Relaxation-Oscillation and Fractional Diffusion-Wave Phenomena, Chaos Soliton. Fract., 7 (1996), 1461–1477. doi: 10.1016/0960-0779(95)00125-5
[18]	R. C. Mittal, R. Rohila, A fourth order cubic B-spline collocation method for the numerical study of the RLW and MRLW equations, Wave Motion, 80 (2018), 47–68. doi: 10.1016/j.wavemoti.2018.04.001
[19]	J. Q. Murillo, S. B. Yuste, An explicit difference method for solving fractional diffusion and diffusion-wave equations in the Caputo form, J. Computat. Nonlin. Dyn., 6 (2011), 021014. doi: 10.1115/1.4002687
[20]	D. A. Murio, Implicit finite difference approximation for time fractional diffusion equations, Comput. Math. Appl., 56 (2008), 1138–1145. doi: 10.1016/j.camwa.2008.02.015
[21]	G. G. O'Brien, M. A. Hyman, S. Kaplan, A study of the numerical solution of partial differential equations, J. Math. Phys., 29 (1950), 223–251. doi: 10.1002/sapm1950291223
[22]	M. Raberto, E. Scalas, F. Mainardi, Waiting-times and returns in high-frequency financial data: an empirical study, Physica A., 314 (2002), 749-755. doi: 10.1016/S0378-4371(02)01048-8
[23]	P. Roul, V. M. K. P. Goura, A high order numerical method and its convergence for time-fractional fourth order partial differential equations, Appl. Math. Comput., 366 (2020), 124727.
[24]	K. Sayevand, A. Yazdani, F. Arjang, Cubic B-spline collocation method and its applicationfor anomalous fractional diffusion equations in transport dynamic systems, J. Vib. Control, 22 (2016), 2173–2186. doi: 10.1177/1077546316636282
[25]	H. Sun, W. Chen, K. Y. Sze, A semi-discrete finite element method for a class of time-fractional diffusion equations, Philos. T. R. Soc. A., 371 (2013), 1–15.
[26]	N. H. Sweilam, M. M. Khader, A. M. S. Mahdy, Crank-Nicolson finite difference method for solving time-fractional diffusion equation, Journal of Fractional Calculus and Applications, 2 (2012), 1–9.
[27]	C. Tadjeran, M. M. Meerschaert, H. Scheffler, A second-order accurate numerical approximation for the fractional diffusion equation, J. Comput. Phys., 213 (2006), 205–213. doi: 10.1016/j.jcp.2005.08.008
[28]	H. S. Shukla, M. Tamsir, An exponential cubic B-spline algorithm for multi-dimensional convection-diffusion equations, Alex. Eng. J., 57 (2018), 1999–2006. doi: 10.1016/j.aej.2017.04.011
[29]	H. S. Shukla, M. Tamsir, R. Jiwari, V. K. Srivastava, A numerical algorithm for computation modeling of 3D nonlinear wave equations based on exponential modified cubic B-spline differential quadrature method, Int. J. Comput. Math., 95 (2017), 752–766.
[30]	M. Tamsir, V. K. Srivastava, N. Dhiman, A. Chauhan, Numerical Computation of Nonlinear Fisher's Reaction–Diffusion Equation with Exponential Modified Cubic B-Spline Differential Quadrature Method, Int. J. Appl. Comput. Math, 4 (2018), 1–13.
[31]	A. Verma, R. Jiwari, M. Koksal, Analytic and numerical solutions of nonlinear diffusion equations via symmetry reductions, Adv. Differ. Equ., 2014 (2014), 1–13. doi: 10.1186/1687-1847-2014-1
[32]	P. Zhuang, F. Liu, Implicit difference approximation for the time fractional diffusion equation, J. Appl. Math. Comput., 22 (2006), 87–99. doi: 10.1007/BF02832039
[33]	L. Wu, S. Zhai, A new high order ADI numerical difference formula for time-fractional convection-diffusion equation, Appl. Math. Comput., 387 (2020), 124564.
[34]	S. Zhai, Z. Weng, X. Feng, J. Yuan, Investigations on several high-order ADI methods for timespace fractional diffusion equation, Nume. Algorithms, 82 (2019), 69-106. doi: 10.1007/s11075-018-0594-z
[35]	S. Zhai, X. Feng, A block-centered finite-difference method for the time-fractional diffusion equation on nonuniform grids, Numer. Heat Tr. B-Fund., 69 (2016), 217–233.
[36]	S. Zhai, Z. Weng, D. Gui, X. Feng, High-order compact operator splitting method for three-dimensional fractional equation with subdiffusion, Int. J. Heat Mass Tran., 84 (2015), 440–447.
[37]	S. Zhai, X. Feng, Y. He, An unconditionally stable compact ADI method for three-dimensional time-fractional convection–diffusion equation, J. Comput. Phys., 269 (2014), 138–155.
[38]	S. Zhai, D. Gui, P. Huang, X. Feng, A novel high-order ADI method for 3D fractionalconvection–diffusion equations, Int. Commun. Heat Mass, 66 (2015), 212–217.
[39]	D. Baleanu, J. A. T. Machado, A. C. Luo, Fractional Dynamics and Control, Springer, 2012.
[40]	M. Caputo, Elasticita e Dissipazione, Zani-Chelli, Bologna, 1969.
[41]	A. Carpinteri, F. Mainardi, Fractals and fractional calculus in Continuum mechanics, Springer Verlag Wien, 1997.
[42]	R. Hilfer, Applications of fractional calculus in physics, World scientific, Singapore, 2000.
[43]	A. Carpinteri, F. Mainardi, Fractals and Fractional Calculus Continuum Mechanics, Springer Verlag Wien, 1997.
[44]	K. S. Miller, B. Ross, An Introduction to the Fractional Calculus and Fractional Differential Equations, Wiley-Interscience, 1993.
[45]	K. B. Oldham, J. Spanier, The Fractional Calculus, Academic Press, New York, 1974.
[46]	I. Podlubny, Fractional differential equations, Academic Press, San Diego, 1999.

This article has been cited by:

1.	Neeraj Dhiman, Mohammad Tamsir, Anand Chauhan, Deependra Nigam, An implicit collocation algorithm based on cubic extended B-splines for Caputo time-fractional PDE, 2021, 22147853, 10.1016/j.matpr.2021.02.230
2.	A.H. Msmali, Mohammad Tamsir, Abdullah Ali H. Ahmadini, Crank-Nicolson-DQM based on cubic exponential B-splines for the approximation of nonlinear Sine-Gordon equation, 2021, 20904479, 10.1016/j.asej.2021.04.004
3.	Fouad Mohammad Salama, Nur Nadiah Abd Hamid, Norhashidah Hj. Mohd Ali, Umair Ali, An efficient modified hybrid explicit group iterative method for the time-fractional diffusion equation in two space dimensions, 2022, 7, 2473-6988, 2370, 10.3934/math.2022134
4.	Majeed A. Yousif, Faraidun K. Hamasalh, A new numerical scheme non-polynomial spline for solving generalized time fractional Fisher equation, 2023, 10641246, 1, 10.3233/JIFS-222445
5.	Numerical Solution of Diffusion Equation with Caputo Time Fractional Derivatives Using Finite-Difference Method with Neumann and Robin Boundary Conditions, 2022, 2663-7804, 95, 10.34104/ijmms.022.0950108
6.	Nabil Korichi, Abdelmalek Boumali, Yassine Chargui, Statistical Properties of the 1D Space Fractional Klein–Gordon Oscillator, 2022, 206, 0022-2291, 32, 10.1007/s10909-021-02638-z
7.	Numerical Solution of Diffusion Equation with Caputo Time Fractional Derivatives Using Finite-Difference Method with Neumann and Robin Boundary Conditions, 2022, 2707-4625, 1, 10.34104/ijmms.022.010014
8.	Younis A Sabawi, Hoshman Q Hamad, Numerical solution of the Whitham-Broer-Kaup shallow water equation by quartic B-spline collocation method, 2024, 99, 0031-8949, 015242, 10.1088/1402-4896/ad1561
9.	Majeed A. Yousif, Faraidun K. Hamasalh, Ahmad Zeeshan, Mohamed Abdelwahed, Muhammad Mubashir Bhatti, Efficient simulation of Time-Fractional Korteweg-de Vries equation via conformable-Caputo non-Polynomial spline method, 2024, 19, 1932-6203, e0303760, 10.1371/journal.pone.0303760
10.	Majeed A. Yousif, Faraidun K. Hamasalh, The fractional non-polynomial spline method: Precision and modeling improvements, 2024, 218, 03784754, 512, 10.1016/j.matcom.2023.11.033
11.	Majeed A. Yousif, Faraidun K. Hamasalh, A Hybrid Non-Polynomial Spline Method and Conformable Fractional Continuity Equation, 2023, 11, 2227-7390, 3799, 10.3390/math11173799
12.	Majeed A. Yousif, Faraidun K. Hamasalh, Conformable non-polynomial spline method: A robust and accurate numerical technique, 2024, 15, 20904479, 102415, 10.1016/j.asej.2023.102415
13.	Mohammad Tamsir, Deependra Nigam, Neeraj Dhiman, Anand Chauhan, A hybrid B-spline collocation technique for the Caputo time fractional nonlinear Burgers’ equation, 2023, 12, 2314-8543, 10.1186/s43088-023-00434-0
14.	Abdelouahed Kouibia, Miguel Pasadas, Loubna Omri, A Shape-Preserving Variational Spline Approximation Problem for Hole Filling in Generalized Offset Surfaces, 2024, 12, 2227-7390, 1736, 10.3390/math12111736
15.	Nazek A. Obeidat, Mahmoud S. Rawashdeh, Omar M. Ababneh, On solving caputo and Riemann–Liouville types of sequential fractional differential equations with boundary conditions using an accurate scheme, 2025, 26668181, 101174, 10.1016/j.padiff.2025.101174

Reader Comments

Your name:*

Email:*
© 2021 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

AIMS Mathematics

1.8 3.4

Metrics

Article views(3897) PDF downloads(314) Cited by(15)

Preview PDF

Download XML

Export Citation

Article outline

Show full outline

Figures and Tables

Figures(4) / Tables(4)

AIMS Mathematics

Approximation of Caputo time-fractional diffusion equation using redefined cubic exponential B-spline collocation technique

Related Papers:

Abstract

1. Introduction

2. Discretization of the problem

3. Stability analysis

4. Computational results and discussions

4.1. Example 1

4.2. Example 2

4.3. Example 3

5. Conclusions

Conflict of interest

References

This article has been cited by:

Reader Comments

通讯作者: 陈斌, bchen63@163.com

Metrics

Figures and Tables

Other Articles By Authors

Catalog

AIMS Mathematics

Approximation of Caputo time-fractional diffusion equation using redefined cubic exponential B-spline collocation technique

Related Papers:

Abstract

1. Introduction

2. Discretization of the problem

3. Stability analysis

4. Computational results and discussions

4.1. Example 1

4.2. Example 2

4.3. Example 3

5. Conclusions

Conflict of interest

References

This article has been cited by:

Reader Comments

通讯作者: 陈斌, bchen63@163.com

Metrics

Figures and Tables

Other Articles By Authors

Related pages

Tools

Export File

Citation

Format

Content

Catalog