Relative perturbation results for eigenvalues and eigenvectors of diagonalisable matrices |
| |
Authors: | S C Eisenstat I C F Ipsen |
| |
Institution: | (1) Department of Computer Science, Yale University, P.O. Box 208285, 06520-8285 New Haven, CT, USA;(2) Department of Mathematics, North Carolina State University, P.O. Box 8205, 27695-8205 Raleigh, NC, USA |
| |
Abstract: | Let
and
be a perturbed eigenpair of a diagonalisable matrixA. The problem is to bound the error in
and
. We present one absolute perturbation bound and two relative perturbation bounds.
The absolute perturbation bound is an extension of Davis and Kahan's sin θ Theorem from Hermitian to diagonalisable matrices.
The two relative perturbation bounds assume that
and
are an exact eigenpair of a perturbed matrixD
1
AD
2
, whereD
1 andD
2 are non-singular, butD
1
AD
2 is not necessarily diagonalisable. We derive a bound on the relative error in
and a sin θ theorem based on a relative eigenvalue separation. The perturbation bounds contain both the deviation ofD
1 andD
2 from similarity and the deviation ofD
2 from identity.
This work was partially supported by NSF grant CCR-9400921. |
| |
Keywords: | 15A18 15A42 15A60 65F15 65F35 65G99 |
本文献已被 SpringerLink 等数据库收录! |
|