K-Spectral Sets
Anne Greenbaum, Natalie Wellen

TL;DR
This paper explores K-spectral sets using spectral set theory, compares bounds derived from different methods, and discusses their effectiveness and limitations in various applications.
Contribution
It introduces new bounds for K-spectral sets based on spectral set theory and compares them with traditional bounds from the Cauchy integral formula.
Findings
New upper bounds on K-values can be tighter than traditional bounds.
In many cases, the bounds from both methods are of similar magnitude.
The bounds from the Cauchy integral formula can be slightly smaller in some scenarios.
Abstract
We use results in [M. Crouzeix and A. Greenbaum,Spectral sets: numerical range and beyond, SIAM Jour. Matrix Anal. Appl., 40 (2019), pp. 1087-1101] to derive a variety of K-spectral sets and show how they can be used in some applications. We compare the K-values derived here to those that can be derived from a straightforward application of the Cauchy integral formula, by replacing the norm of the integral by the integral of the resolvent norm. While, in some cases, the new upper bounds on the optimal K-value are much tighter than those from the Cauchy integral formula, we show that in many cases of interest, the two values are of the same order of magnitude, with the bounds from the Cauchy integral formula actually being slightly smaller. We give a partial explanation of this in terms of the numerical range of the resolvent at points near an ill-conditioned eigenvalue.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsMatrix Theory and Algorithms · Advanced Optimization Algorithms Research
