Numerische Mathematik

, Volume 13, Issue 4, pp 316–322 | Cite as

On comparatively stable tridiagonalization methods

  • John W. Rainey
Article

Abstract

A method for choosing certain matrices necessary for the “tridiagonalization” of an arbitrary, real, square matrix is sketched. As opposed to previous methods which choose modifications of matrices developed for other purposes, we choose these matrices on the basis of a direct examination of the essential computational problem occuring in “sequential tridiagonalization” and require that our choices have relatively small condition numbers.

Keywords

Mathematical Method Condition Number Previous Method Small Condition Computational Problem 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Bauer, F. L.: Sequential reduction to tridiagonal form. J. SIAM7, 107–113 (1959).Google Scholar
  2. 2.
    Householder, A. S., andF. L. Bauer: On certain methods for expanding the characteristic polynomial. Num. Math.1, 29–37 (1959)Google Scholar
  3. 3.
    ——: The theory of matrices in numerical analysis. New York: Blaisdell 1964.Google Scholar
  4. 4.
    Labudde, C. D.: The reduction of an arbitrary square matrix to tridiagonal form using similarity transformations. Math. Com.17, 433–437 (1963)Google Scholar
  5. 5.
    Lambert, R. J., andW. L. Waltman: T-algorithm for tridiagonalization. J. SIAM13, 1069–1078 (1965)Google Scholar
  6. 6.
    Lanczos, C.: An iterative method for the solution of the eigenvalue problem of linear differential and integral operators. J. Res. Nat. Bur. Standards45, 255–282 (1950).Google Scholar
  7. 7.
    RAINEY, J. W., andG. J. HABETLER: An application of the LR factorization to sequential tridiagonalization methods. SIAM J. Appl. Math. 17 (1969) (to Appear).Google Scholar
  8. 8.
    Wilkinson, J. H.: The algebraic eigenvalue problem. Oxford-New York 1965Google Scholar

Copyright information

© Springer-Verlag 1969

Authors and Affiliations

  • John W. Rainey
    • 1
  1. 1.Department of MathematicsUniversity of TulsaTulsaUSA

Personalised recommendations