Müntz Spectral Method for Two-Dimensional Space-Fractional Convection-Diffusion Equation

Communications in Computational Physics
Vol. 26 No. 5 (2019), pp. 1415-1443
Preview Buy Article · USD $20 Full text requires purchase or subscription
Author(s)
, ,
1 School of Mathematics and Statistics, Jiangsu Normal University, 221116 Xuzhou, China.
2 Jiangsu Normal Univ, Sch Math & Stat, Xuzhou 221116, Jiangsu, Peoples R China
3 Xiamen Univ, Sch Math Sci, Xiamen 361005, Fujian, Peoples R China
4 Xiamen Univ, Fujian Prov Key Lab Math Modeling & High Performa, Xiamen 361005, Fujian, Peoples R China
5 Bordeaux Institut National Polytechnique, I2M UMR 5295, France
6 Bordeaux INP, Lab 12M UMR 5295, F-33607 Pessac, France
7 School of Mathematical Sciences, Xiamen University, 361005 Xiamen, China
Received
February 28, 2019
Accepted
May 7, 2019
Abstract

In this paper, we propose and analyze a Müntz spectral method for a class of two-dimensional space-fractional convection-diffusion equations. The proposed methods make new use of the fractional polynomials, also known as Müntz polynomials, which can be regarded as continuation of our previous work. The extension is twofold. Firstly, the existing Müntz spectral method for fractional differential equation with fractional derivative order 0<µ<1 is generalized to 0<µ≤2, which is nontrivial since the classical Müntz polynomials only have no more than H1/2+λ regularity, where 0 < λ ≤ 1 is the characteristic parameter of the Müntz polynomial. Secondly, 1D Müntz spectral method is extended to the 2D space-fractional convection-diffusion equation. Compared to the time-fractional diffusion equation, some new operators such as suitable H1-projectors are needed to analyze the error of the numerical solution. The main contribution of the present paper consists of an efficient method combining the Crank-Nicolson scheme for the temporal discretization and a new spectral method using the Müntz Jacobi polynomials for the spatial discretization of the 2D space-fractional convection-diffusion equation. A detailed convergence analysis is carried out, and several error estimates are established. Finally, a series of numerical experiments is performed to verify the theoretical claims.

Stay updated
Share
How to Cite