Advertisement

Planar-CG Methods and Matrix Tridiagonalization in Large Scale Unconstrained Optimization

  • Giovanni Fasano
Part of the Applied Optimization book series (APOP, volume 82)

Abstract

In this paper we aim at carrying out and describing some issues for real eigenvalue computation via iterative methods. More specifically we work out new techniques for iteratively developing specific tridiagonalizations of a symmetric and indefinite matrix AR n × n , by means of suitable Krylov subspace algorithms defined in [16], [26]. These schemes represent extensions of the well known Conjugate Gradient (CG) method to the indefinite case. We briefly recall these algorithms and we suggest a comparison with the method in [22], along with a discussion on the practical application of the proposed results for eigenvalue computation. Furthermore, we focus on motivating the fruitful use of these tridiagonalizations for ensuring the convergence to second order points, within an optimization framework.

Keywords

unconstrained optimization eigenvalue computation matrix tridiagonalization Conjugate Gradient Krylov subspace methods 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    G.S.Ammar, L.Reichel, D.C.Sorensen (1992), “An implementation of a divide and conquer algorithm for the unitary eigenproblem”, ACM Transactions on Mathematical Software,vol. 18, pp. 292 – 307.zbMATHCrossRefGoogle Scholar
  2. [2]
    W.Barth, R.S.Martin, J.H.Wilkinson (1967), “Calculation of the eigenvalues of a Symmetric Tridiagonal Matrix by the Method of Bisection”, Numerische Mathematik, vol. 9, pp. 379–404.MathSciNetCrossRefGoogle Scholar
  3. [3]
    C.H.Bischof, M.Marques, X.Sun (1993), “Parallel Bandreduction and Tridiagonalization”, Proceedings of the Sixth SIAM Conference on Parallel Processing for Scientific Computing, R.Sincovec, Ed., pp.389–390, SIAM.Google Scholar
  4. [4]
    C.H.Bischof, X.Sun (1992), “A Framework for Symmetric Band Reduction and Tridiagonalization”, Preprint MCS-P298–0392, Mathematics and Computer Science Division, Argonne National Laboratory, Argonne, IL 60439–4801.Google Scholar
  5. [5]
    C.H.Bischof, X.Sun (1995), “On tridiagonalizing and Diagonalizing Symmetric Matrices with Repeated Eigenvalues”, Argonne Preprint MCS-P545–1095, PRISM Working Note #25, Mathematics and Computer Science Division, Argonne National Laboratory, Argonne, IL 60439–4801.Google Scholar
  6. [6]
    H.Bowdler, R.S.Martin, C.Reinsch, J.H.Wilkinson (1968), “The QR and QL Algorithms for Symmetric Matrices”, Numerische Mathematik, vol. 11, pp. 293–306.MathSciNetzbMATHCrossRefGoogle Scholar
  7. [7]
    W.W.Bradbury, R.Fletcher (1966), “New iterative methods for solutions of the eigenproblem”, Numerische Mathematik, vol. 9, pp. 259–267.MathSciNetzbMATHCrossRefGoogle Scholar
  8. [8]
    J.J.M.Cuppen (1981), “A divide and conquer method for the symmetric tridiagonal eigenproblem”, Numer. Math., vol. 36, pp. 177–195.MathSciNetzbMATHCrossRefGoogle Scholar
  9. [9]
    R.S.Dembo, T.Steihaug (1983), “Truncated-Newton Algorithms for large scale Matrix Methods”, Mathematical Programming, vol. 26, pp. 190–212.MathSciNetzbMATHCrossRefGoogle Scholar
  10. [10]
    J.W.Demmel (1997), Applied Numerical Linear Algebra, SIAM, Philadelphia.zbMATHGoogle Scholar
  11. [11]
    J.Demmel, K.Veselic ’ (1992), “Jacobi’s method is more accurate than QR”, SIAM J. Matrix Anal. Appl.,vol. 13, pp. 1204–1246 (LAPACK Working Note 15).MathSciNetzbMATHCrossRefGoogle Scholar
  12. [12]
    A.Edelman, T.Arias, S.T.Smith (1994), “Curvature in Conjugate Gradient Eigenvalue Computation with Applications to Materials and Chemistry Calculations”, Proceedings of the SIAM Applied Linear Algebra Conference, J.G.Lewis, ed., SIAM, Philadelphia, pp. 233–238.Google Scholar
  13. [13]
    A.Edelman, T.Arias, S.T.Smith (to appear), “The geometry of algorithms with orthogonality constraints”, SIAM J. Matrix Anal. Appl.Google Scholar
  14. [14]
    A.Edelman, S.T.Smith (1996), “On Conjugate Gradient-like methods for eigen-like problems”, BIT, vol. 36:1, pp. 494–508.MathSciNetzbMATHCrossRefGoogle Scholar
  15. [15]
    G.Fasano (2002), “On Some Properties of Planar-CG algorithms for Large Scale Unconstrained Optimization - Part A” T.R. 03–02, Dipartimento di Informatica e Sistemistica ‘A.Ruberti’, Università “La Sapienza” Roma, Italy.Google Scholar
  16. [16]
    G.Fasano (2001), “A new CG-based method for the solution of large scale indefinite linear systems”, T.R. 08–01, Dipartimento di Informatica e Sistemistica ‘A.Ruberti’, Università “La Sapienza” Roma, Italy.Google Scholar
  17. [17]
    W.N.Gansterer, D.F.Kvasnicka, C.W.Ueberhuber (1998), “Numerical Experiments with Symmetric Eigensolvers”, AURORA TR1998–19, University of Technology, Vienna.Google Scholar
  18. [18]
    G.H.Golub, C.F.Van Loan (1989), Matrix computations - 3rd edition, The John Hopkins Press, Baltimore.Google Scholar
  19. [19]
    N.I.M.Gould, S.Lucidi, M.Roma, Ph.L.Toint (1999), “Solving the trust-region subproblem using the Lanczos method”, SIAM Journal on Optimization vol. 9, pp. 504–525.MathSciNetzbMATHCrossRefGoogle Scholar
  20. [20]
    N.I.M.Gould, S.Lucidi, M.Roma, Ph.L.Toint (2000), “Exploiting Negative Curvature Directions in Linesearch Methods for Unconstrained Optimization”, Optimization Methods and Software vol. 14, pp. 75–98.MathSciNetzbMATHCrossRefGoogle Scholar
  21. [21]
    M.R.Hestenes (1980), Conjugate Direction Methods in Optimization, Springer Verlag, New York Heidelberg Berlin.zbMATHGoogle Scholar
  22. [22]
    C.Lanczos (1950), “An Iterative Method for the Solution of the Eigenvalue Problem of Linear Differential and Integral Operators”, Journal of Research of the National Bureau of Standards vol. 45, Research Paper 2133.Google Scholar
  23. [23]
    B.Lang (1999), “Out-of-Core Solution of Large Symmetric Eigenproblems”, Preprint RWTH-CS-SC-99–03, Aachen University of Technology, Institute for Scientific Computing.Google Scholar
  24. [24]
    S.Lucidi, M.Roma (1997), “Numerical experiences with new truncated Newton methods in large scale unconstrained optimization”, Computational Optimization and Applications, vol. 7, pp. 71–87.MathSciNetzbMATHCrossRefGoogle Scholar
  25. [25]
    S.Lucidi, F.Rochetich, M.Roma (1999), “Curvilinear stabilization techniques for Truncated Newton methods in large scale unconstrained Optimization”, SIAM Journal on Optimization,vol. 8, pp. 916–939.MathSciNetCrossRefGoogle Scholar
  26. [26]
    D.G.Luenberger (1969), “Hyperbolic pairs in the Method of Conjugate Gradients”, SIAM J. Appl. Math.,vol. 17, pp. 1263–1267.MathSciNetzbMATHCrossRefGoogle Scholar
  27. [27]
    M.Mongeau, M.Torki (1999), “Computing eigenelements of real symmetric matrices via optimization”, T.R. MIP 99 – 54.Google Scholar
  28. [28]
    J.J.More, D.C.Sorensen (1979), “On the use of directions of negative curvature in a modified Newton method”, Mathematical Programming,vol. 16, pp. 1–20.MathSciNetzbMATHCrossRefGoogle Scholar
  29. [29]
    C.C.Paige, M.A.Saunders (1975), “Solution of sparse indefinite systems of linear equations”, SIAM J. on Numerical Analysis,vol. 12, pp. 617 – 629.MathSciNetzbMATHCrossRefGoogle Scholar
  30. [30]
    B.Parlett (1980), The symmetric eigenvalue problem, Prentice-Hall series in Computational Mathematics, Englewood Cliffs.zbMATHGoogle Scholar
  31. [31]
    B.Parlett (1974), “Generalized Rayleigh Methods with Applications to Finding Eigenvalues of Large Matrices”, Lin. Alg. and its Applic.,vol. 4, pp. 353–368.Google Scholar
  32. [32]
    J.Stoer (1983), Mathematical Programming, The State of the Art,A.Bachem, M.Grotschel, B.Korte eds., Springer-Verlag, Berlin.Google Scholar
  33. [33]
    H.Van Der Vorst (1996), “Subspace Iteration for Eigenproblems” CWI Quarterly, vol. 9, pp. 151–160.MathSciNetzbMATHGoogle Scholar
  34. [34]
    J.H.Wilkinson (1965), The Algebraic Eigenvalue Problem, Oxford: Oxford University (Clarendon).zbMATHGoogle Scholar

Copyright information

© Kluwer Academic Publishers B.V. 2003

Authors and Affiliations

  • Giovanni Fasano
    • 1
  1. 1.Dipartimento di Informatica e Sistemistica “Antonio Ruberti”Università di Roma “La Sapienza”RomaItaly

Personalised recommendations