Theoretical and computable optimal subspace expansions for matrix eigenvalue problems

DOI10.1137/20M1331032MaRDI QIDQ5071436zbMATH OpenOpenAlexFDO

Authors Zhongxiao Jia

Publication date 21 April 2022

Published in SIAM Journal on Matrix Analysis and Applications (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2004.04928

zbMATH Keywords

computable optimally expanded subspaceeigenvalue problemharmonic Ritz vectornon-Krylov subspaceoptimal expansion vectoroptimal subspace expansionrefined harmonic Ritz vectorrefined Ritz vectorRitz vector

Mathematics Subject Classification ID

Eigenvalues, singular values, and eigenvectors (15A18) Numerical computation of eigenvalues and eigenvectors of matrices (65F15)

Abstract: Consider the optimal subspace expansion problem for the matrix eigenvalue problem $Ax=lambdax$: Which vector $w$ in the current subspace $mathcalV$, after multiplied by $A$, provides an optimal subspace expansion for approximating a desired eigenvector $x$ in the sense that $x$ has the smallest angle with the expanded subspace $mathcal{V}_w=mathcalV+mspanAw$, i.e., $w_{opt}=argma{x}_{winmathcalV}cosangle(mathcal{V}_w,x)$? This problem is important as many iterative methods construct nested subspaces that successively expand $mathcalV$ to $mathcal{V}_w$. An expression of $w_{opt}$ by Ye (Linear Algebra Appl., 428 (2008), pp. 911--918) for $A$ general, but it could not be exploited to construct a computable (nearly) optimally expanded subspace. He turns to deriving a maximization characterization of $cosangle(mathcal{V}_w,x)$ for a {em given} $winmathcalV$ when $A$ is Hermitian. We generalize Ye's maximization characterization to the general case and find its maximizer. Our main contributions consist of explicit expressions of $w_{opt}$, $(I-P_V)A{w}_{opt}$ and the optimally expanded subspace $mathcal{V}_{w_{opt}}$ for $A$ general, where $P_V$ is the orthogonal projector onto $mathcalV$. These results are fully exploited to obtain computable optimally expanded subspaces within the framework of the standard, harmonic, refined, and refined harmonic Rayleigh--Ritz methods. We show how to efficiently implement the proposed subspace expansion approaches. Numerical experiments demonstrate the effectiveness of our computable optimal expansions.

Recommendations

• Optimal expansion of subspaces for eigenvector approximations
• Generalized subspace iteration method for solving matrix pair eigenproblem
• On expansion of search subspaces for large non-Hermitian eigenproblems
• A subspace method for large-scale eigenvalue optimization
• Efficient expansion of subspaces in the Jacobi-Davidson method for standard and generalized eigenproblems

Cites work

• A Jacobi–Davidson Iteration Method for Linear Eigenvalue Problems
• A new justification of the Jacobi-Davidson method for large eigenproblems
• An analysis of the Rayleigh-Ritz method for approximating eigenspaces
• Computational methods for large eigenvalue problems
• Efficient Algorithms for Computing a Strong Rank-Revealing QR Factorization
• Harmonic and refined harmonic shift-invert residual Arnoldi and Jacobi-Davidson methods for interior eigenvalue problems
• scientific article; zbMATH DE number 47363 (Why is no real title available?)
• scientific article; zbMATH DE number 1049353 (Why is no real title available?)
• scientific article; zbMATH DE number 6159604 (Why is no real title available?)
• Inner iterations in the shift-invert residual Arnoldi method and the Jacobi-Davidson method
• Matrix algorithms. Vol. 2: Eigensystems
• Numerical methods for large eigenvalue problems
• On expansion of search subspaces for large non-Hermitian eigenproblems
• Optimal expansion of subspaces for eigenvector approximations
• Refined iterative algorithms based on Arnoldi's process for large unsymmetric eigenproblems
• Some theoretical comparisons of refined Ritz vectors and Ritz vectors
• Templates for the Solution of Algebraic Eigenvalue Problems
• The Convergence of Generalized Lanczos Methods for Large Unsymmetric Eigenproblems
• The convergence of harmonic Ritz values, harmonic Ritz vectors and refined harmonic Ritz vectors
• The convergence of harmonic Ritz vectors and harmonic Ritz values, revisited
• The Lanczos and Conjugate Gradient Algorithms
• Theory of Inexact Krylov Subspace Methods and Applications to Scientific Computing
• Variable Accuracy of Matrix-Vector Products in Projection Methods for Eigencomputation

Cited in

(3)

• On expansion of search subspaces for large non-Hermitian eigenproblems
• Optimal expansion of subspaces for eigenvector approximations
• Subspace acceleration for the Crawford number and related eigenvalue optimization problems

Describes a project that uses

Uses Software

• JDQR
• JDQZ

This page was built for publication: Theoretical and computable optimal subspace expansions for matrix eigenvalue problems

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5071436)