A Statistical Evaluation of Inverse Iteration

  • E. R. Jessup
Conference paper
Part of the NATO ASI Series book series (volume 70)

Abstract

The eigenvalues of a symmetric tridiagonal matrix can be computed by bisection using Sturm sequences and their corresponding eigenvectors found by inverse iteration. The independent computing tasks comprising this combination are especially appropriate for implementation on a distributed memory multiprocessor, but the parallel efficiency of bisection and inverse iteration is not enough to recommend their use in general. Bisection produces eigenvalues to high accuracy. Inverse iteration results in eigenvectors with quality dependent on the spacing of the eigenvalues and on the starting vector. This paper examines the factors influencing the accuracy of inverse iteration. The effects of random starting vectors are discussed, and some uses of statistical analysis in the design of an inverse iteration algorithm are presented.

Keywords

Verse 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    M. Abramowitz and I.A. Stegun, Handbook of Mathematical Functions, Dover Publications, Inc., 1965.Google Scholar
  2. [2]
    H. Bowdler, R.S. Martin, and J.H. Wilkinson, The QR and QL Algorithms for Symmetric Matrices, Num. Math., 11 (1968), pp. 227–240.MathSciNetCrossRefGoogle Scholar
  3. [3]
    J.J.M. Cuppen, A Divide and Conquer Method for the Symmetric Tridiagonal Eigenproblem, Numer. Math., 36 (1981), pp. 177–95.MathSciNetCrossRefMATHGoogle Scholar
  4. [4]
    J. Demmel and W. Kahan, LAPACK Working Note#3: Computing Small Singular Values of Bidiagonal Matrices with Guaranteed Relative Accuracy, Mathematics and Computer Science Division, Argonne National Laboratory, 1988.Google Scholar
  5. [5]
    L. Devroye, Non-Uniform Random Variate Generation, Springer-Verlag, 1986.Google Scholar
  6. [6]
    J.D. Dixon, Estimating Extremal Eigenvalues and Condition Numbers of Matrices, SIAM J. Numer. Anal., 20 (1983), pp. 812–814.MathSciNetCrossRefMATHGoogle Scholar
  7. [7]
    I.C.F. Ipsen and E.R. Jessup, Solving the Symmetric Tridiagonal Eigenvalue Problem on the Hypercube, Research Report 548, Dept. Computer Science, Yale University, 1987.Google Scholar
  8. [8]
    R.V. Jensen and R. Shankar, Statistical Behavior in Deterministic Quantum Systems with Few Degrees of Freedom, Phys. Rev. Lett., 54 (1985), pp. 1879–1882.CrossRefGoogle Scholar
  9. [9]
    N.L. Johnson and S. Kotz, Continuous Univariate Distributions, Houghton Mifflin Company, 1970.Google Scholar
  10. [10]
    A.S. Krishnakumar and M. Morf, A Tree Architecture for the Symmetric Eigenproblem, Proc. 27th Annual Symp. of SPIE, 1983.Google Scholar
  11. [11]
    J.R. Kuttler and V.G. Sigillito, Eigenvalues of the Laplacian in Two Dimensions, SIAM Review, 26 (1984), pp. 163–193.MathSciNetCrossRefMATHGoogle Scholar
  12. [12]
    B.N. Parlett, The Symmetric Eigenvalue Problem, Prentice Hall, Englewood Cliffs, NJ, 1980.MATHGoogle Scholar
  13. [13]
    B.T. Smith, J.M. Boyle, J.J. Dongarra, B.S. Garbow, Y. Ikebe, V.C. Klema, and C.B. Moler, Matrix Eigensystem Routines-EISPACK Guide, Lecture Notes in Computer Science, Vol. 6, 2nd edition, Springer-Verlag, 1976.Google Scholar
  14. [14]
    M.D. Springer, The Algebra of Random Variables, John Wiley and Sons, 1979.Google Scholar
  15. [15]
    J.H. Wilkinson, The Algebraic Eigenvalue Problem, Clarendon Press, Oxford, 1965.MATHGoogle Scholar
  16. [16]
    J.H. Wilkinson and G. Peters, The Calculation of Specified Eigen-vectors by Inverse Iteration, Handbook for Automatic Computation, Vol.II: Linear Algebra, Springer Verlag, 1971, pp. 418–439.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1991

Authors and Affiliations

  • E. R. Jessup
    • 1
  1. 1.Department of Computer ScienceYale UniversityNew HavenUSA

Personalised recommendations