Transactions of the Japan Society for Industrial and Applied Mathematics
Online ISSN : 2424-0982
ISSN-L : 0917-2246
A Parameter Estimation for the Preconditioning Using Sparse Direct Solvers for Krylov Subspace Methods Based on the Minimal Residual Approach(Application)
Ikuro YamazakiAkira ImakuraHiroto TadanoTetsuya Sakurai
Author information
JOURNAL FREE ACCESS

2013 Volume 23 Issue 3 Pages 381-404

Details
Abstract
We consider solving the large linear systems involving a matrix which has a relatively large number of small nonzero elements. As suitable preconditionings for Krylov subspace methods to solve such linear systems, the techniques using sparse direct solvers for a sparse approximate coefficient matrix have been proposed. The sparse approximate coefficient matrix is generated by the so-called cutoff, and the performance of the preconditioning is highly dependent on the cutoff parameter. In this paper, we present a technique to suitably estimate a cutoff parameter. The validity of our technique is verified by some numerical experiments.
Content from these authors
© 2013 The Japan Society for Industrial and Applied Mathematics
Previous article Next article
feedback
Top