A comparative study of preconditioned Lanczos methods for nonsymmetric linear system

Technical Report · Tue Sep 01 04:00:00 EDT 1992

OSTI ID:6885222

Tong, C H

Many important scientific applications require the solution of very large and sparse nonsymmetric linear systems. Hence, there has been great interest in the development of robust and efficient solvers. In recent years, the class of Krylov subspace methods has emerged as the most popular iterative solution schemes. Coupled with effective preconditioning, these methods show fast convergence for many problems. However, these methods are not guranteed to converge, especially for many difficult problems that are strongly nonsymmetric, near singular and indefinite. In this report, we survey some recent developments of Krylov subspace methods which are based on the Lanczos procedure. These methods have been implemented and tested on a variety of problems. Extensive numerical results are presented and discussed.

• Research Organization: Sandia National Labs., Livermore, CA (United States)
• Sponsoring Organization: DOE; USDOE, Washington, DC (United States)
• DOE Contract Number: AC04-76DR00789
• OSTI ID: 6885222
• Report Number(s): SAND-91-8240B; ON: DE93006453
• Country of Publication: United States
• Language: English

Related Subjects

99 GENERAL AND MISCELLANEOUS
990200* -- Mathematics & Computers
ALGORITHMS
ASYMMETRY
CALCULATION METHODS
COMPARATIVE EVALUATIONS
CONVERGENCE
EVALUATION
ITERATIVE METHODS
MATHEMATICAL LOGIC
MATRICES