Change of matrix eigenvalues multiplicity associated with perturbation of a diagonal element or deletion of a row and column
 
 
Description:  Given an nxn matrix A, an eigenvalue λ of A and an index i (1\le i \le n) we say that i is downer (respectively neutral or Parter), if the multiplicity of λ in the matrix A(i) (obtained from A by deleting row and column i is smaller (resp equal or greater) than the multiplicity of λ in A. We will gave an equivalent formulation of an index being downer, neutral or Parter in terms of perturbation on the diagonal entry in position A of A. Of course every index must be downer for at least an eigenvalue; in fact, for an irreducible Hermitian matrix A every index must be downer for at least two (distinct) eigenvalues. If the graph of A is a tree we will characterize the trees vertices (indices) that can be downers for just two eigenvalues. We will also discuss the more general problem of characterizing for a given k the trees and vertices that can be downers for k eigenvalues.
Date:  2015-12-16
Start Time:   15:00
Speaker:  Antonio Leal Duarte (CMUC, Univ. Coimbra)
Institution:  CMUC-UC
Place:  Room 5.5 DMUC
Research Groups: -Algebra and Combinatorics
See more:   <Main>  
 
© Centre for Mathematics, University of Coimbra, funded by
Science and Technology Foundation
Powered by: rdOnWeb v1.4 | technical support