A lopsided scaled DTS preconditioning method for the discrete space-fractional diffusion equations. (September 2022)

• Record Type: Journal Article
• Authors: Tang, Shi-Ping; Huang, Yu-Mei
• Is Part Of: Applied mathematics letters. Volume 131(2022)
• Publication Date: 2022-09
• Subjects: Space-fractional diffusion equations -- Lopsided matrix splitting iteration methods -- Preconditioner -- Spectral distribution; Applied mathematics -- Periodicals; 519.05
• DOI: 10.1016/j.aml.2022.108022
• Languages: English
• ISSNs: 0893-9659

Abstract

Abstract: The Grünwald–Letnikov type formulas are applied to discretize the space-fractional diffusion equation and the discretization result is got as a system of linear equations. A lopsided scaled diagonal and Toeplitz splitting (LSDTS) iteration method is constructed to solve this linear system and the convergence is also discussed. An LSDTS τ preconditioner is proposed based on the LSDTS iteration method and the GMRES method combined with this preconditioner is applied to solve the linear system. We theoretically show that the eigenvalues of the LSDTS τ preconditioned matrix are clustered. The numerical experiments illustrate that the LSDTS τ preconditioner can significantly improve the convergence properties of the GMRES method.