Minimal residual method stronger than polynomial preconditioning
- Los Alamos National Lab., NM (United States); and others
Two popular methods for solving symmetric and nonsymmetric systems of equations are the minimal residual method, implemented by algorithms such as GMRES, and polynomial preconditioning methods. In this study results are given on the convergence rates of these methods for various classes of matrices. It is shown that for some matrices, such as normal matrices, the convergence rates for GMRES and for the optimal polynomial preconditioning are the same, and for other matrices such as the upper triangular Toeplitz matrices, it is at least assured that if one method converges then the other must converge. On the other hand, it is shown that matrices exist for which restarted GMRES always converges but any polynomial preconditioning of corresponding degree makes no progress toward the solution for some initial error. The implications of these results for these and other iterative methods are discussed.
- Research Organization:
- Front Range Scientific Computations, Inc., Boulder, CO (United States); US Department of Energy (USDOE), Washington DC (United States); National Science Foundation, Washington, DC (United States)
- OSTI ID:
- 223856
- Report Number(s):
- CONF-9404305-Vol.1; ON: DE96005735; TRN: 96:002320-0030
- Resource Relation:
- Journal Volume: 17; Journal Issue: 4; Conference: Colorado conference on iterative methods, Breckenridge, CO (United States), 5-9 Apr 1994; Other Information: PBD: [1994]; Related Information: Is Part Of Colorado Conference on iterative methods. Volume 1; PB: 203 p.
- Country of Publication:
- United States
- Language:
- English
Similar Records
Polynomial preconditioning for conjugate gradient methods
A comparison of preconditioned Krylov subspace methods for large‐scale nonsymmetric linear systems