论文信息 - Spectral sensitivity of products of bidiagonals

Spectral sensitivity of products of bidiagonals

Abstract Simple formulae are derived for certain relative condition numbers. Tridiagonal matrices may be represented as products of bidiagonals in various ways depending on properties such as symmetry and positive definiteness. The condition numbers give the amplification factor for relative changes in a nonzero eigenvalue caused by relative changes in an entry of a bidiagonal factor. The formulae show that in many, but not all cases these condition numbers are of modest size. Several examples illustrate the results and raise new questions.

Beresford N. Parlett

[1] Z. Drmač,et al. On Scaled Almost-Diagonal Hermitian Matrix Pairs , 1997 .

[2] Ren-Cang Li. Relative Perturbation Theory: I. Eigenvalue and Singular Value Variations , 1998, SIAM J. Matrix Anal. Appl..

[3] J. Barlow,et al. Computing accurate eigensystems of scaled diagonally dominant matrices: LAPACK working note No. 7 , 1988 .

[4] Ilse C. F. Ipsen,et al. Relative perturbation techniques for singular value problems , 1995 .

[5] K. Veselic,et al. On linear vibrational systems with one dimensional damping , 1988 .

[6] James Demmel,et al. Accurate Singular Values of Bidiagonal Matrices , 1990, SIAM J. Sci. Comput..

[7] J. Demmel,et al. The bidiagonal singular value decomposition and Hamiltonian mechanics: LAPACK working note No. 11 , 1989 .

[8] Ivan Slapničar,et al. Floating-point perturbations of Hermitian matrices , 1993 .

[9] J. H. Wilkinson. The algebraic eigenvalue problem , 1966 .

[10] James Demmel,et al. Jacobi's Method is More Accurate than QR , 1989, SIAM J. Matrix Anal. Appl..

[11] B. Parlett. The Symmetric Eigenvalue Problem , 1981 .

[12] Ivan Slapničar,et al. Accurate Symmetric Eigenreduction by a Jacobi Method , 1993 .

[13] B. Noble. Applied Linear Algebra , 1969 .