Spectral Perturbation Bounds for Positive Definite Matrices
Author:
- Roy MathiasRMRoy Mathias
- 1 October 1997
- journal article
- Published by Society for Industrial & Applied Mathematics (SIAM) in SIAM Journal on Matrix Analysis and Applications
- Vol. 18 (4) , 959-980
Abstract
Let H and H + $\Delta$ H be positive definite matrices. It was shown by Barlow and Demmel and Demmel and Veselic that if one takes a componentwise approach one can prove much stronger bounds on $\lambda_i(H)/\lambda_i(H + \Delta H)$ and the components of the eigenvectors of H and H + $\Delta$ H than by using the standard normwise perturbation theory. Here a unified approach is presented that improves on the results of Barlow, Demmel, and Veselic. It is also shown that the growth factor associated with the error bound on the components of the eigenvectors computed by Jacobi's method grows linearly (rather than exponentially) with the number of Jacobi iterations required for convergence.
Keywords
This publication has 11 references indexed in Scilit:
- A relative perturbation bound for positive definite matricesLinear Algebra and its Applications, 1998
- A relative perturbation bound for positive definite matricesLinear Algebra and its Applications, 1998
- A Bound for the Matrix Square Root with Application to Eigenvector PerturbationSIAM Journal on Matrix Analysis and Applications, 1997
- Fast accurate eigenvalue methods for graded positive definite matricesNumerische Mathematik, 1996
- Relative Perturbation Techniques for Singular Value ProblemsSIAM Journal on Numerical Analysis, 1995
- The Instability of Parallel Prefix Matrix MultiplicationSIAM Journal on Scientific Computing, 1995
- Accurate Eigensystem Computations by Jacobi MethodsSIAM Journal on Matrix Analysis and Applications, 1995
- A note on Jacobi Being More Accurate Than $QR$SIAM Journal on Matrix Analysis and Applications, 1994
- Jacobi’s Method is More Accurate than QRSIAM Journal on Matrix Analysis and Applications, 1992
- Computing Accurate Eigensystems of Scaled Diagonally Dominant MatricesSIAM Journal on Numerical Analysis, 1990