Abstract
Lanczos' method for solving the system of linear equationsAx=b consists in constructing a sequence of vectors (x k ) such thatr k =b−Ax k =P k (A)r 0 wherer 0=b−Ax 0.P k is an orthogonal polynomial which is computed recursively. The conjugate gradient squared algorithm (CGS) consists in takingr k =P 2k (A)r0. In the recurrence relation forP k , the coefficients are given as ratios of scalar products. When a scalar product in a denominator is zero, then a breakdown occurs in the algorithm. When such a scalar product is close to zero, then rounding errors can seriously affect the algorithm, a situation known as near-breakdown. In this paper it is shown how to avoid near-breakdown in the CGS algorithm in order to obtain a more stable method.
Similar content being viewed by others
References
R.E. Bank and T.F. Chan, A composite step bi-conjugate gradient algorithm for solving nonsymmetric systems, Numer. Algorithms, this issue.
R.E. Bank and T.F. Chan, An analysis of the composite step bi-conjugate gradient algorithm for solving nonsymmetric systems, Numer. Math. 66 (1993) 295–319.
C. Brezinski,Padé-Type Approximation and General Orthogonal Polynomials, ISNM vol. 50 (Birkhäser, Basel, 1980).
C. Brezinksi and M. Redivo-Zaglia,Extrapolation Methods. Theory and Practice (North-Holland, Amsterdam, 1991).
C. Brezinski and M. Redivo-Zaglia, A new presentation of orthogonal polynomials with applications to their computation. Numer. Algorithms 1 (1991) 207–221.
C. Brezinski and M. Redivo-Zaglia, Treatment of near-breakdown in the CGS algorithm, Publication ANO 257, Université des Sciences et Technologies de Lille (November 1991).
C. Brezinski and M. Redivo-Zaglia, Hybrid procedures for solving linear systems, Numer. Math. 67 (1994) 1–19.
C. Brezinski and M. Redivo-Zaglia, Breakdowns in the computation of orthogonal polynomials, in:Nonlinear Numerical Methods and Rational Approximation, ed. A. Cuyt, (Kluwer, Dordrecht, 1994) pp. 49–59.
C. Brezinski and M. Redivo-Zaglia, Look-ahead in Bi-CGSTAB and other methods for linear systems, to appear.
C. Brezinski, M. Redivo-Zaglia and H. Sadok, Avoiding breakdown and near-breakdown in Lanczos type algorithms, Numer. Algorithms 1 (1991) 261–284.
C. Brezinski, M. Redivo-Zaglia and H. Sadok, Addendum to “Avoiding breakdown and near-breakdown in Lanczos type algorithms”, Numer. Algorithms 2 (1992) 133–136.
C. Brezinski, M. Redivo-Zaglia and H. Sadok, A breakdown-free Lanczos type algorithm for solving linear systems, Numer. Math. 63 (1992) 29–38.
C. Brezinski, M. Redivo-Zaglia and H. Sadok, Breakdowns in the implementation of the Lánczos method for solving linear systems, Comp. Math. Appl., to appear.
C. Brezinski and H. Sadok, Lanczos type methods for solving systems of linear equations, Appl. Numer. Math. 11 (1993) 443–473.
C. Brezinski and H. Sadok, Avoiding breakdown in the CGS algorithm, Numer. Algorithms 1 (1991) 199–206.
P.N. Brown, A theoretical comparison of the Arnoldi and GMRES algorithms, SIAM J. Sci. Stat. Comp. 12 (1991) 58–78.
T.F. Chan and T. Szeto, A composite step conjugate gradients squared algorithm for solving nonsymmetric linear systems, Numer. Algorithms, this issue.
J.M. Chesneaux, Study of the computing accuracy by using probabilistic approach, in:Contribution to Computer Arithmetic and Self-Validating Numerical Methods, ed. C. Ulrich (Baltzer, Basel, 1990) pp. 19–30.
J.M. Chesneaux, Stochastic arithmetic properties, in:Computational and Applied Mathematics, I, eds. C. Brezinski and U. Kulisch (North-Holland, Amsterdam, 1992), pp. 81–91.
J.M. Chesneaux and J. Vignes, Les fondements de l'arithmétique stochastique, C.R. Acad. Sci. Paris, I, 315 (1992) 1435–1440.
J.M. Chesneaux and J. Vignes, L'algorithme de Gauss en arithmétique stochastique, C. R. Acad. Sci. Paris, II, 316 (1993) 171–176.
A. Draux,Polynômes Orthogonaux Formels. Applications, LNM 974 (Springer, Berlin, 1983).
R. Fletcher, Conjugate gradient methods for indefinite systems, in:Numerical Analysis, ed. G.A. Watson, LNM 506 (Springer, Berlin, 1976) pp. 73–89.
R.W. Freund, M.H. Gutknecht and N.M. Nachtigal, An implementation of the look-ahead Lanczos algorithm for non-Hermitian matrices, SIAM J. Sci. Comp. 14 (1993) 137–158.
M.H. Gutknecht, A completed theory of the unsymmetric Lanczos process and related algorithms, Part I, SIAM J. Matrix. Anal. Appl. 13 (1992) 594–639.
M.H. Gutknecht, The unsymmetric Lanczos algorithms and their relations to Padé approximation, continued fractions and the qd algorithm, to appear.
M.H. Gutknecht, Variants of BICGSTAB for matrices with complex spectrum, SIAM J. Sci. Comp. 14 (1993) 1020–1033.
K.C. Jea and D.M. Young, On the simplification of generalized conjugate gradient methods for nonsymmetrizable linear systems, Lin. Alg. Appl. 52/53 (1983) 399–417.
W. Joubert, Generalized conjugate gradient and Lanczos methods for the solution of nonsymmetric systems of linear equations, Ph.D. Thesis, University of Texas at Austin, Austin (1990).
T.W. Körner,Fourier Analysis (Cambridge University Press, Cambridge, 1988).
C. Lanczos, An iteration method for the solution of the eigenvalue problem of linear differential and integral operators. J. Res. Natl. Bur. Stand. 45 (1950) 255–282.
C. Lanczos, Solution of systems of linear equations by minimized iterations, J. Res. Natl. Bur. Stand. 49 (1952) 33–53.
N.M. Nachtigal, S.C. Reddy and L.N. Trefethen, How fast are nonsymmetric matrix iterations?, SIAM J. Matrix Anal. Appl. 13 (1992) 778–795.
B.N. Parlett, D.R. Taylor and Z.A. Liu, A look-ahead Lanczos algorithm for unsymmetric matrices, Math. Comp. 44 (1985) 105–124.
W. Schönauer,Scientific Computing on Vector Computers (North-Holland, Amsterdam, 1987).
P. Sonneveld, CGS, a fast Lanczos-type solver for nonsymmetric linear systems, SIAM J. Sci. Stat. Comp. 10 (1989) 36–52.
H.A. Van der Vorst, The convergence behavior of preconditioned CG and CG-S, in:Preconditioned Conjugate Gradient Methods, eds. O. Axelsson and L.Yu. Kolotilina, LNM 1457 (Springer, Berlin, 1990) pp. 126–136.
H.A. Van der Vorst, BI-CGSTAB: a fast and smoothly converging variant of Bi-CG for the solution of nonsymmetric linear systems, SIAM J. Sci. Stat. Comp. 13 (1992) 631–644.
J. Vignes, Review of stochastic approach to round-off error analysis and its applications, Math. Comp. Simul. 30 (1988) 481–491.
P.K.W. Vinsome, Orthomin, an iterative method for solving sparse sets of simultaneous linear equations, in:Proc. 4th Symp. on Reservoir Simulation (Society of Petroleum Engineers of AIME, 1976), pp. 149–159.
R. Weiss, Convergence behavior of generalized conjugate gradient methods, Thesis, University of Karlsruhe (1990).
D.M. Young and K.C. Jea, Generalized conjugate-gradient acceleration of nonsymmetrizable iterative methods, Lin. Alg. Appl. 34 (1980) 159–194.
L. Zhou and H.F. Walker, Residual smoothing techniques for iterative methods, SIAM J. Sci. Stat. Comp., to appear.
R. Zippel,Effective Polynomial Computation, (Kluwer, Dordrecht, 1993).
Author information
Authors and Affiliations
Additional information
Communicated by T.F. Chan
Rights and permissions
About this article
Cite this article
Brezinski, C., Redivo-Zaglia, M. Treatment of near-breakdown in the CGS algorithm. Numer Algor 7, 33–73 (1994). https://doi.org/10.1007/BF02141260
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF02141260