A preconditioner based on a splitting-type iteration method for solving complex symmetric indefinite linear systems
| Autor: | Yu-Tao Zheng, Lu-Bin Cui, Xiao-Qing Zhang |
|---|---|
| Rok vydání: | 2021 |
| Předmět: |
Iterative method
Preconditioner Applied Mathematics Linear system General Engineering 010103 numerical & computational mathematics Krylov subspace Computer Science::Numerical Analysis 01 natural sciences Generalized minimal residual method Mathematics::Numerical Analysis 010101 applied mathematics Convergence (routing) Computer Science::Mathematical Software Applied mathematics 0101 mathematics Coefficient matrix Eigenvalues and eigenvectors Mathematics |
| Zdroj: | Japan Journal of Industrial and Applied Mathematics. 38:965-978 |
| ISSN: | 1868-937X 0916-7005 |
| DOI: | 10.1007/s13160-021-00471-1 |
| Popis: | In this paper, we propose a preconditioned modified positive/negative-stable splitting (PMPNS) iteration method to solve complex symmetric indefinite linear system more efficiently. By analyzing the convergence of the PMPNS iteration method and discussing the spectral properties of the PMPNS iteration method, we construct a new preconditioner to make the eigenvalues of the coefficient matrix more aggregated, which leads to fast convergence of Krylov subspace iteration methods such as GMRES. Numerical example is given to illustrate the efficiency of the PMPNS preconditioner used in GMRES method. In particular, the GMRES method with the PMPNS preconditioner demonstrates meshsize-independent convergence behavior. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink