scientific article; zbMATH DE number 1569645

MaRDI QIDQ4762997zbMATH OpenFDO

Authors

Publication date 26 February 2001

zbMATH Keywords

convergenceefficiency estimatoreigenvalueinverse iterationsMarkov chainMonte Carlo algorithmsMonte Carlo iterationsnumerical examplesparallel algorithmparallel computationresolvent matrix

Mathematics Subject Classification ID

Monte Carlo methods (65C05) Numerical computation of eigenvalues and eigenvectors of matrices (65F15) Parallel numerical computation (65Y05)

Review text

The problem of evaluating the smallest eigenvalue of a real symmetric matrix using statistical numerical methods is considered. Two new almost optimal Monte Carlo algorithms are presented: A resolvent Monte Carlo algorithm (RMC) which uses Monte Carlo iterations by the resolvent matrix and includes a parameter controlling the rate of convergence and a Monte Carlo algorithm with inverse iterations. Numerical tests are performed for a number of large sparse symmetric test matrices. The tests are performed on supercomputers Intel-PARAGON (which is a distributed memory multicomputer) and CRAY Y-MP C92A (a two-processor vector machine). Some information about the parallel efficiency of the algorithms is obtained, showing that the algorithms under consideration are well-parallized and well-vectorizable.

Recommendations

• Parallel computations of eigenvalues based on a Monte Carlo approach
• Monte Carlo algorithms for calculating eigenvalues
• A Parallel Inertia Method for Finding Eigenvalues on Vector and SIMD Architectures
• Implementation of Monte Carlo algorithms for eigenvalue problem using MPI
• Computational Science – ICCS 2005

Cited in

(5)

• Parallel stochastic estimation method of eigenvalue distribution
• New hybrid Monte Carlo methods and computing the dominant generalized eigenvalue
• Implementation of Monte Carlo algorithms for eigenvalue problem using MPI
• Parallel computations of eigenvalues based on a Monte Carlo approach
• Computational Science – ICCS 2005

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4762997)