A bound on the joint spectral radius using the diagonals

Publication date 1 December 2020

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

Abstract: We present a combinatorial approach to studying the joint spectral radius by the following formula: Given a finite set

S i g m a

of nonnegative matrices, we can express the joint spectral radius

h o (S i g m a)

by [ ho(Sigma)=sup_nmax_i sqrt[n]{max_{A_1,dots,A_ninSigma} (A_1dots A_n)_{i,i}}. ] Corollaries of the approach include a simple proof of the joint spectral radius theorem for finite sets of nonnegative matrices and the following theoretical bound on

h o (S i g m a)

: There exists some constant

K

so that [ sqrt[m_i]{max_i max_{A_1,dots,A_{m_i}inSigma} (A_1dots A_{m_i})_{i,i}} le ho(Sigma) le sqrt[m_i]{K max_i max_{A_1,dots,A_{m_i}inSigma} (A_1dots A_{m_i})_{i,i}} ] where

m_{i}

for each

i

is any number so that there exist

A_{1}, d o t s, A_{m_{i}} i n S i g m a

satisfying

(A_{1} d o t s A_{m_{i}})_{i, i} > 0

, or

m_{i} = 1

if there are no such matrices. Some of the results in the work are not completely new with some known related forms, but the merits of the approach lie in the simplicity of the proofs and its combinatorial nature.

This page was built for publication: A bound on the joint spectral radius using the diagonals

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6355036)