QRT: A QR-Based Tridiagonalization Algorithm for Nonsymmetric Matrices

The stable similarity reduction of a nonsymmetric square matrix to tridiagonal form has been a long-standing problem in numerical linear algebra. The biorthogonal Lanczos process is in principle a candidate method for this task, but in practice it is confined to sparse matrices and is restarted periodically because roundoff errors affect its three-term recurrence scheme and degrade the biorthogonality after a few steps. This adds to its vulnerability to serious breakdowns or near-breakdowns, the handling of which involves recovery strategies such as the look-ahead technique, which needs a careful implementation to produce a block-tridiagonal form with unpredictable block sizes. Other candidate methods, geared generally towards full matrices, rely on elementary similarity transformations that are prone to numerical instabilities. Such concomitant difficulties have hampered finding a satisfactory solution to the problem for either sparse or full matrices. This study focuses primarily on full matrices. After outlining earlier tridiagonalization algorithms from within a general framework, we present a new elimination technique combining orthogonal similarity transformations that are stable. We also discuss heuristics to circumvent breakdowns. Applications of this study include eigenvalue calculation and the approximation of matrix functions.

[1]  Jorgen L. Nikolajsen.  An Improved Laguerre Eigensolver for Unsymmetric Matrices , 2000, SIAM J. Sci. Comput..

[2]  George A. Geist.  Reduction of a general matrix to tridiagonal form using a hypercube multiprocessor , 1991 .

[3]  C. Donald LaBudde,et al.  The reduction of an arbitrary real square matrix to tridiagonal form using similarity transformations , 1963 .

[4]  Christopher Strachey,et al.  The Reduction of a Matrix to Codiagonal Form by Eliminations , 1961, Comput. J..

[5]  Alan George,et al.  A Finite Procedure for the Tridiagonalization of a General Matrix , 1995, SIAM J. Matrix Anal. Appl..

[6]  R. T. Gregory,et al.  On the reduction of an arbitrary real square matrix to tridiagonal form , 1964 .

[7]  Roger B. Sidje,et al.  Expokit: a software package for computing matrix exponentials , 1998, TOMS.

[8]  Beresford N. Parlett.  A note on La Budde’s algorithm , 1964 .

[9]  Roland W. Freund,et al.  An Implementation of the Look-Ahead Lanczos Algorithm for Non-Hermitian Matrices , 1993, SIAM J. Sci. Comput..

[10]  S. Kaniel,et al.  The ELR Method for Computing the Eigenvalues of a General Matrix , 1981 .

[11]  Jack J. Dongarra,et al.  Algorithm 710: FORTRAN subroutines for computing the eigenvalues and eigenvectors of a general matrix by reduction to general tridiagonal form , 1990, TOMS.

[12]  F. L. Bauer.  Sequential Reduction to Tridiagonal Form , 1959 .

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

[14]  Gary W. Howell,et al.  Efficient computation of eigenvalues of randomly generated matrices , 1994 .

[15]  Roger B. Sidje,et al.  An Augmented Lanczos Algorithm for the Efficient Computation of a Dot-Product of a Function of a Large Sparse Symmetric Matrix , 2003, International Conference on Computational Science.

[16]  Nicholas J. Higham,et al.  A Schur-Parlett Algorithm for Computing Matrix Functions , 2003, SIAM J. Matrix Anal. Appl..

[17]  Beresford N. Parlett,et al.  Reduction to Tridiagonal Form and Minimal Realizations , 1992, SIAM J. Matrix Anal. Appl..