Computable eigenvalue bounds for rank-\(k\) perturbations (Q968983): Difference between revisions

The authors investigate lower bounds for the eigenvalues of low rank Hermitian perturbations of Hermitian matrices and for the singular values of low rank perturbations of general matrices. They show that, for the smallest eigenvalue, some earlier results of Weyl and of \textit{I. C. F. Ipsen} and \textit{B. Nadler} [SIAM J. Matrix Anal. Appl. 31, 40--53 (2009; Zbl 1189.15022)] are the first two of a sequence of increasingly tight bounds. Generally the \(q\)th element of the sequence is an eigenvalue of a \(q\times q\) matrix.