An augmented matrix-based CJ-FEAST SVDsolver for computing a partial singular value decomposition with the singular values in a given interval

TL;DR

This paper introduces an augmented matrix-based CJ-FEAST SVDsolver that improves accuracy and stability in computing partial singular value decompositions within a specified interval, outperforming previous methods.

Contribution

The paper develops a new augmented matrix-based CJ-FEAST SVDsolver that enhances accuracy, stability, and convergence for partial SVD computations in a specified spectrum interval.

Findings

The new solver is always numerically backward stable.

It provides accurate singular triplet approximations within the interval.

Numerical experiments confirm improved robustness and efficiency.

Abstract

The cross-product matrix-based CJ-FEAST SVDsolver proposed previously by the authors is shown to compute the left singular vector possibly much less accurately than the right singular vector and may be numerically backward unstable when a desired singular value is small. In this paper, an alternative augmented matrix-based CJ-FEAST SVDsolver is considered to compute the singular triplets of a large matrix $A$ with the singular values in an interval $[a,b]$ contained in the singular spectrum. The new CJ-FEAST SVDsolver is a subspace iteration applied to an approximate spectral projector of the augmented matrix $[0, A^T; A, 0]$ associated with the eigenvalues in $[a,b]$, and constructs approximate left and right singular subspaces with the desired singular values independently, onto which $A$ is projected to obtain the Ritz approximations to the desired singular triplets. Compact estimates are given for the accuracy of the approximate spectral projector, and a number of convergence results are established. The new solver is proved to be always numerically backward stable. A convergence comparison of the cross-product and augmented matrix-based CJ-FEAST SVDsolvers is made, and a general-purpose choice strategy between the two solvers is proposed for the robustness and overall efficiency. Numerical experiments confirm all the results.