Benzi, Michele, Evans, Thomas M., Hamilton, Steven P., Lupo Pasini, Massimiliano, and Slattery, Stuart R.
Subjects
ITERATIVE methods (Mathematics), LINEAR systems, MONTE Carlo method, SPARSE approximations, NUMERICAL analysis
Abstract
We consider hybrid deterministic-stochastic iterative algorithms for the solution of large, sparse linear systems. Starting from a convergent splitting of the coefficient matrix, we analyze various types of Monte Carlo acceleration schemes applied to the original preconditioned Richardson (stationary) iteration. These methods are expected to have considerable potential for resiliency to faults when implemented on massively parallel machines. We establish sufficient conditions for the convergence of the hybrid schemes, and we investigate different types of preconditioners including sparse approximate inverses. Numerical experiments on linear systems arising from the discretization of partial differential equations are presented. [ABSTRACT FROM AUTHOR]
Summary: For large sparse non‐Hermitian positive definite linear systems, we establish exact and inexact quasi‐HSS iteration methods and discuss their convergence properties. Numerical experiments show that both iteration methods are effective and robust when they are used either as linear solvers or as matrix splitting preconditioners for the Krylov subspace iteration methods. In addition, these two iteration methods are, respectively, much more powerful than the exact and inexact HSS iteration methods, especially when the linear systems have nearly singular Hermitian parts or strongly dominant skew‐Hermitian parts, and they can be employed to solve non‐Hermitian indefinite linear systems with only mild indefiniteness. [ABSTRACT FROM AUTHOR]