An Implementation of the QMR Method Based on Coupled Two-Term Recurrences

Authors: Roland W. Freund and Noël M. NachtigalAuthors Info & Affiliations

https://doi.org/10.1137/0915022

Tools

Abstract

Recently, the authors proposed a new Krylov subspace iteration, the quasi-minimal residual (QMR) algorithm, for solving non-Hermitian linear systems. In the original implementation of the QMR method, the Lanczos process with look-ahead is used to generate basis vectors for the underlying Krylov subspaces. In the Lanczos algorithm, these basis vectors are computed by means of three-term recurrences. It has been observed that, in finite-precision arithmetic, vector iterations based on three-term recursions are usually less robust than mathematically equivalent coupled two-term vector recurrences.

This paper presents a look-ahead algorithm that constructs the Lanczos basis vectors by means of coupled two-term recursions. Some implementation details are given, and the look-ahead strategy is described. A new implementation of the QMR method, based on this coupled two-term algorithm, is proposed. A simplified version of the QMR algorithm without look-ahead is also presented, and the special case of QMR for complex symmetric linear systems is considered. Results of numerical experiments comparing the original and the new implementations of the QMR method are reported.

MSC codes

Keywords

Get full access to this article

View all available purchase options and get full access to this article.

References

1.

J. K. Cullum, R. A. Willoughby, J. K. Cullum, R. A. Willoughby, A practical procedure for computing eigenvalues of large sparse nonsymmetric matricesLarge scale eigenvalue problems (Oberlech, 1985), North-Holland Math. Stud., Vol. 127, North-Holland, Amsterdam, 1986, 193–240

2.

André Draux, Polynômes orthogonaux formels, Lecture Notes in Mathematics, Vol. 974, Springer-Verlag, Berlin, 1983vi+625

3.

I. S. Duff, R. G. Grimes, J. G. Lewis, Sparse matrix test problems, ACM Trans. Math. Software, 15 (1989), 1–14

4.

R. Fletcher, G. A. Watson, Conjugate gradient methods for indefinite systemsNumerical analysis (Proc 6th Biennial Dundee Conf., Univ. Dundee, Dundee, 1975), Springer, Berlin, 1976, 73–89. Lecture Notes in Math., Vol. 506

5.

Roland W. Freund, Conjugate gradient-type methods for linear systems with complex symmetric coefficient matrices, SIAM J. Sci. Statist. Comput., 13 (1992), 425–448

6.

Roland W. Freund, D. Braess, L. L. Schumaker, Quasi-kernel polynomials and convergence results for quasi-minimal residual iterationsNumerical methods in approximation theory, Vol. 9 (Oberwolfach, 1991), Internat. Ser. Numer. Math., Vol. 105, Birkhäuser, Basel, 1992, 77–95

7.

Roland W. Freund, Martin H. Gutknecht, Noël M. Nachtigal, An implementation of the look-ahead Lanczos algorithm for non-Hermitian matrices, SIAM J. Sci. Comput., 14 (1993), 137–158

8.

R. W. Freund, M. Hochbruck, On the use of two QMR algorithms for solving singular systems and applications in Markov chain modeling, Tech. Report, 91.25, RIACS, NASA Ames Research Center, Moffett Field, CA, 1991, Dec.

9.

Roland W. Freund, Noël M. Nachtigal, QMR: a quasi-minimal residual method for non-Hermitian linear systems, Numer. Math., 60 (1991), 315–339

10.

R. W. Freund, N. M. Nachtigal, An implementation of the QMR method based on coupled two-term recurrences, Tech. Report, 92.15, RIACS, NASA Ames Research Center, Moffett Field, CA, 1992, June

11.

R. W. Freund, T. Szeto, A quasi-minimal residual squared algorithm for non-Hermitian linear systems, Proc. 1992 Copper Mountain Conf. on Iterative Methods, 1992, April

12.

M. H. Gutknecht, A completed theory of the unsymmetric Lanczos process and related algorithms, Part II, IPS Research Report, 90–16, IPS, ETH, Zürich, Switzerland, 1990, Sept.

13.

Martin H. Gutknecht, A completed theory of the unsymmetric Lanczos process and related algorithms. I, SIAM J. Matrix Anal. Appl., 13 (1992), 594–639

14.

Magnus R. Hestenes, Eduard Stiefel, Methods of conjugate gradients for solving linear systems, J. Research Nat. Bur. Standards, 49 (1952), 409–436 (1953)

15.

W. D. Joubert, Ph.D. Thesis, Generalized Conjugate Gradient and Lanczos Methods for the Solution of Nonsymmetric Systems of Linear Equations, Center for Numerical Analysis, The University of Texas at Austin, 1990, Jan.

16.

Cornelius Lanczos, An iteration method for the solution of the eigenvalue problem of linear differential and integral operators, J. Research Nat. Bur. Standards, 45 (1950), 255–282

17.

Cornelius Lanczos, Solution of systems of linear equations by minimized-iterations, J. Research Nat. Bur. Standards, 49 (1952), 33–53

18.

Hans Petter Langtangen, Aslak Tveito, A numerical comparison of conjugate gradient-like methods, Comm. Appl. Numer. Methods, 4 (1988), 793–798

19.

Beresford N. Parlett, Derek R. Taylor, Zhishun A. Liu, A look-ahead Lánczos algorithm for unsymmetric matrices, Math. Comp., 44 (1985), 105–124

20.

J. K. Reid, J. K. Reid, On the method of conjugate gradients for the solution of large sparse systems of linear equationsLarge sparse sets of linear equations (Proc. Conf., St. Catherine's Coll., Oxford, 1970), Academic Press, London, 1971, 231–254

21.

Heinz Rutishauser, Der Quotienten-Differenzen-Algorithmus, Mitt. Inst. Angew. Math. Zürich, 1957 (1957), 74–, an der ETH, E. Stiefel, ed., Birkhaüser, Basel

22.

Y. Saad, ILUT: a dual threshold incomplete LU factorization, Research Report, UMSI 92/38, University of Minnesota Supercomputer Institute, Minneapolis, 1992, March

23.

D. R. Taylor, Ph.D. Thesis, Analysis of the Look Ahead Lanczos Algorithm, Dept. of Mathematics, University of California, Berkeley, CA, 1982, Nov.

24.

H. Woźniakowski, Roundoff-error analysis of a new class of conjugate-gradient algorithms, Linear Algebra Appl., 29 (1980), 507–529