Fernando's solution to Wilkinson's problem: An application of double factorization

Abstract Suppose that one knows a very accurate approximation σ to an eigenvalue λ of a symmetric tridiagonal matrix T . A good way to approximate the eigenvector x is to discard an appropriate equation, say the rth , from the system (T − σI) x = 0 and then to solve the resulting underdetermined system in any of several stable ways. However the output x can be completely inaccurate if r is chosen poorly, and in the absence of a quick and reliable way to choose r , this method has lain neglected for over 35 years. Experts in boundary value problems have known about the special structure of the inverse of a tridiagonal matrix since the 1960s, and their double triangular factorization technique (down and up) gives directly the redundancy of each equation and so reveals the set of good choices for r . The relation of double factorization to the eigenvector algorithm of Godunov and his collaborators is described. The results extend to band matrices and to zero entries in eigenvectors, and have uses beyond eigenvector computation.

[1]  S. Godunov.  Guaranteed Accuracy in Numerical Linear Algebra , 1993 .

[2]  W. Givens.  Numerical Computation of the Characteristic Values of a Real Symmetric Matrix , 1954 .

[3]  Wallace Givens,et al.  The Characteristic Value-Vector Problem , 1957, JACM.

[4]  Gene H. Golub,et al.  On Fourier-Toeplitz methods for separable elliptic problems , 1974 .

[5]  Gérard Meurant,et al.  A Review on the Inverse of Symmetric Tridiagonal and Block Tridiagonal Matrices , 1992, SIAM J. Matrix Anal. Appl..

[6]  K. V. Fernando.  On computing an eigenvector of a tridiagonal matrix , 1997 .

[7]  Ilse C. F. Ipsen.  A history of inverse iteration , 1994 .

[8]  Gilbert Strang,et al.  Implicit difference methods for initial-boundary value problems☆ , 1966 .

[9]  James Hardy Wilkinson,et al.  The Calculation of the Eigenvectors of Codiagonal Matrices , 1958, Comput. J..

[10]  G. TEMPLE,et al.  Relaxation Methods in Engineering Science , 1942, Nature.

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

[12]  Henk A. van der Vorst.  Analysis of a parallel solution method for tridiagonal linear systems , 1987, Parallel Comput..

[13]  V. I. Kostin,et al.  Computation of an eigenvector of a symmetric tridiagonal matrix , 1985 .

[14]  K. V. Fernando.  Accurate BABE Factorisation of Tridiagonal Matrices for Eigenproblems , 1995 .

[15]  I. Babuvška.  Numerical stability in problems of linear algebra. , 1972 .

[16]  Peter Henrici.  Bounds for Eigenvalues of Certain Tridiagonal Matrices , 1963 .

[17]  Jessica Engel,et al.  Problem , 1902 .

[18]  Ed Anderson,et al.  LAPACK users' guide - [release 1.0] , 1992 .

[19]  Beresford N. Parlett,et al.  Invariant subspaces for tightly clustered eigenvalues of tridiagonals , 1996 .

[20]  Ilse C. F. Ipsen,et al.  Improving the Accuracy of Inverse Iteration , 1992, SIAM J. Sci. Comput..

[21]  N. S. Barnett,et al.  Private communication , 1969 .

[22]  I. Dhillon.  Reliable Computation of the Condition Number of a Tridiagonal Matrix in O(n) Time , 1998, SIAM J. Matrix Anal. Appl..

[23]  N. Higham.  Efficient algorithms for computing the condition number of a tridagonal matrix , 1986 .

[24]  D. Kershaw.  Inequalities on the elements of the inverse of a certain tridiagonal matrix , 1970 .

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