English

Fast Computation of Sep$_\lambda$ via Interpolation-based Globality Certificates

Optimization and Control 2023-01-24 v5 Numerical Analysis Numerical Analysis

Abstract

Given two square matrices AA and BB, we propose a new approach for computing the smallest value ε0\varepsilon \geq 0 such that A+EA+E and A+FA+F share an eigenvalue, where E=F=ε\|E\|=\|F\|=\varepsilon. In 2006, Gu and Overton proposed the first algorithm for computing this quantity, called sepλ(A,B)\mathrm{sep}_\lambda(A,B) ("sep-lambda"), using ideas inspired from an earlier algorithm of Gu for computing the distance to uncontrollability. However, the algorithm of Gu and Overton is extremely expensive, which limits it to the tiniest of problems, and until now, no other algorithms have been known. Our new algorithm can be orders of magnitude faster and can solve problems where AA and BB are of moderate size. Moreover, our method consists of many "embarrassingly parallel" computations, and so it can be further accelerated on multi-core hardware. Finally, we also propose the first algorithm to compute an earlier version of sep-lambda where E+F=ε\|E\| + \|F\|=\varepsilon.

Keywords

Cite

@article{arxiv.1911.05136,
  title  = {Fast Computation of Sep$_\lambda$ via Interpolation-based Globality Certificates},
  author = {Tim Mitchell},
  journal= {arXiv preprint arXiv:1911.05136},
  year   = {2023}
}

Comments

4th Revision, January 21, 2023

R2 v1 2026-06-23T12:13:34.854Z