Scaled Companion Linearizations for Coefficientwise Backward Stability

DOI10.48550/arXiv.2610.20285MaRDI QIDQ7395257zbMATH OpenarXivFDO

Authors Javier Perez, Javier Pérez

Publication date 14 October 2026

Published in arXiv

Tldr We construct a family of scaled companion linearizations that cover the nonzero complex plane. Each is paired with a region and eigenvector recovery rule, such that backward-error amplification factors are bounded at exact pencil eigenvectors, independently of coefficient norm disparities.

Full work available at URL https://doi.org/10.48550/arXiv.2610.20285

Mathematics Subject Classification ID

Numerical computation of eigenvalues and eigenvectors of matrices (65F15) Numerical computation of matrix norms, conditioning, scaling (65F35) Roundoff error (65G50)

arXiv classification

math.NA

Cites work

• A backward stable algorithm for quadratic eigenvalue problems
• A simplified approach to Fiedler-like pencils via block minimal bases pencils
• An Algorithm for Generalized Matrix Eigenvalue Problems
• An algorithm for the complete solution of quadratic eigenvalue problems
• Backward error and condition of polynomial eigenvalue problems
• Backward Error of Polynomial Eigenproblems Solved by Linearization
• Block Kronecker Linearizations of Matrix Polynomials and their Backward Errors
• Fiedler companion linearizations and the recovery of minimal indices
• NLEVP, a collection of nonlinear eigenvalue problems
• Polynomial eigenvalue solver based on tropically scaled Lagrange linearization
• Tropical roots as approximations to eigenvalues of matrix polynomials
• Tropical scaling of polynomial matrices
• Vector Spaces of Linearizations for Matrix Polynomials
• Vector spaces of linearizations for matrix polynomials: a bivariate polynomial approach

This page was built for publication: Scaled Companion Linearizations for Coefficientwise Backward Stability

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7395257)