A note on weighted FOM and GMRES for solving nonsymmetric linear systems

被引：7

作者：

Cao, ZH ^{[1
]}

Yu, XY

机构：

[1] Fudan Univ, Dept Math, Shanghai 200433, Peoples R China

[2] Fudan Univ, Lab Math Nonlinear Sci, Shanghai 200433, Peoples R China

来源：

APPLIED MATHEMATICS AND COMPUTATION | 2004年 / 151卷 / 03期

基金：

中国国家自然科学基金;

关键词：

nonsymmetric linear systems; sparse matrix; iterative methods; preconditioning; Krylov subspaces; Arnoldi; FOM; GMRES;

D O I：

10.1016/S0096-3003(03)00373-4

中图分类号：

O29 [应用数学];

学科分类号：

070104 ;

摘要：

Recently Essai [Numer. Algorithm 18 (1998) 277-292] presented two new methods called WFOM and WGMRES for solving nonsymmetric linear systems. In this note we first point out a scaling invariant property of these methods, then we discuss the performance of the preconditioned weighted FOM and GMRES. Experimental results are presented to show that, contrast to WFOM and WGMRES, the preconditioned weighed FOM and GMRES have not so good performance compared to preconditioned FOM(m) and preconditioned GMRES(m). (C) 2003 Elsevier Inc. All rights reserved.

引用

页码：719 / 727

页数：9