On the convergence of restarted Krylov subspace methods (Q2706266)
From MaRDI portal
 | This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use this page instead for the normal view: On the convergence of restarted Krylov subspace methods |
scientific article
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | On the convergence of restarted Krylov subspace methods |
scientific article |
Statements
19 March 2001
0 references
GMRES method
0 references
restarted iterative method
0 references
full orthogonal method
0 references
FOM methods
0 references
Krylov subspace iterative method
0 references
convergence
0 references
large nonsymmetric linear systems
0 references
performance
0 references
numerical experiments
0 references
minimum perturbation methods
0 references
On the convergence of restarted Krylov subspace methods (English)
0 references
The paper is concerned with investigation of convergence of Krylov subspace iterative method for the solution of large nonsymmetric linear systems. Restarted methods terminate the process after a fixed number of iterations and then repeat the procedure using the residual of the current approximate solution as new initial vector. NEWLINENEWLINENEWLINEThe aim of this paper is to analyze the computation of the restarted quantities in order to identify the quantities that influence the performance degradation of the restarted methods and to compare the performance of different restarted methods in the given setting. The author has highlighted the quantities that play an important role at restart time and has given closed forms for the approximate solutions of restarted generalized minimal residual (GMRES) and full orthogonal (FOM) methods. Numerical experiments show that restarted FOM may be superior to the GMRES and minimum perturbation methods.
0 references
