VPAStab(J,L): An iterative method with look-ahead for the solution of large sparse linear systems

Authors:
P. R. Graves-Morris
Affiliations:
Department of Computing, University of Bradford, Bradford BD7 1DP, UK
Venue:
Journal of Computational and Applied Mathematics
Year:
2009

Citing 10
Cited 0

BI-CGSTAB: a fast and smoothly converging variant of BI-CG for the solution of nonsymmetric linear systems

SIAM Journal on Scientific and Statistical Computing
A weakly stable algorithm for Pade´ approximants and the inversion of Hankel matrices

SIAM Journal on Matrix Analysis and Applications
Variants of BICGSTAB for matrices with complex spectrum

SIAM Journal on Scientific Computing
Reliable updated residuals in hybrid Bi-CG methods

Computing
Composite Step Product Methods for Solving Nonsymmetric Linear Systems

SIAM Journal on Scientific Computing
Estimating the Attainable Accuracy of Recursively Computed Residual Methods

SIAM Journal on Matrix Analysis and Applications
Look-Ahead Procedures for Lanczos-Type Product Methods Based on Three-Term Lanczos Recurrences

SIAM Journal on Matrix Analysis and Applications
Numerical Linear Algebra for High Performance Computers

Numerical Linear Algebra for High Performance Computers
Accuracy and Stability of Numerical Algorithms

Accuracy and Stability of Numerical Algorithms
BiCGStab, VPAStab and an adaptation to mildly nonlinear systems

Journal of Computational and Applied Mathematics

Quantified Score

Hi-index	7.29

Visualization

Abstract

VPAStab(J,L) denotes an iterative method designed to give an accurate numerical solution of a large sparse system of linear equations. The iteration comprises L-steps and possibly jump-steps. An L-step involves stabilisation (minimisation) over an L-dimensional subspace following L levels of vector-Pade approximation. A jump-step (look-ahead stage) avoids near-breakdowns by jumping past them by J levels simultaneously. A numerical implementation of VPAStab(2,L) is considered in the context of some well-known examples.