Extrapolating the Arnoldi algorithm to improve eigenvector convergence

MaRDI QIDQ5035918zbMATH OpenFDO

Authors Sara Pollock, L. Ridgway Scott

Publication date 22 February 2022

Full work available at URL https://arxiv.org/abs/2103.08635, https://www.global-sci.org/intro/article_detail/ijnam/19389.html

zbMATH Keywords

Arnoldi algorithmeigenvalue computationextrapolation

Mathematics Subject Classification ID

Extrapolation to the limit, deferred corrections (65B05) Numerical computation of eigenvalues and eigenvectors of matrices (65F15)

Abstract: We consider extrapolation of the Arnoldi algorithm to accelerate computation of the dominant eigenvalue/eigenvector pair. The basic algorithm uses sequences of Krylov vectors to form a small eigenproblem which is solved exactly. The two dominant eigenvectors output from consecutive Arnoldi steps are then recombined to form an extrapolated iterate, and this accelerated iterate is used to restart the next Arnoldi process. We present numerical results testing the algorithm on a variety of cases and find on most examples it substantially improves the performance of restarted Arnoldi. The extrapolation is a simple post-processing step which has minimal computational cost.

Recommendations

• Publication:4503865
• An Arnoldi-extrapolation algorithm for computing pagerank
• A new method for accelerating Arnoldi algorithms for large scale eigenproblems
• Vector extrapolation methods with applications to solution of large systems of equations and to PageRank computations
• scientific article; zbMATH DE number 16693

Cites work

• A simple extrapolation method for clustered eigenvalues
• Deflation Techniques for an Implicitly Restarted Arnoldi Iteration
• Modified Gram-Schmidt (MGS), Least Squares, and Backward Stability of MGS-GMRES
• Numerical methods for large eigenvalue problems
• Numerics of Gram-Schmidt orthogonalization
• On the Rates of Convergence of the Lanczos and the Block-Lanczos Methods
• SLEPc
• The University of Florida sparse matrix collection
• Toward the optimal preconditioned eigensolver: Locally optimal block preconditioned conjugate gradient method
• Van der Waals interactions between two hydrogen atoms: the Slater-Kirkwood method revisited
• Variations on Arnoldi's method for computing eigenelements of large unsymmetric matrices

Cited in

(3)

• A simple extrapolation method for clustered eigenvalues
• scientific article; zbMATH DE number 1507134 (Why is no real title available?)
• Accelerating the computation of tensor Z-eigenvalues

Describes a project that uses

Uses Software

• SLEPc
• SparseMatrix
• lobpcg.m

This page was built for publication: Extrapolating the Arnoldi algorithm to improve eigenvector convergence

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5035918)