Advertisement

BIT Numerical Mathematics

, Volume 35, Issue 4, pp 588–604 | Cite as

Solution of sparse rectangular systems using LSQR and CRAIG

  • Michael A. Saunders
Article

Abstract

We examine two iterative methods for solving rectangular systems of linear equations: LSQR for over-determined systemsAx ≈ b, and Craig's method for under-determined systemsAx = b. By including regularization, we extend Craig's method to incompatible systems, and observe that it solves the same damped least-squares problems as LSQR. The methods may therefore be compared on rectangular systems of arbitrary shape.

Various methods for symmetric and unsymmetric systems are reviewed to illustrate the parallels. We see that the extension of Craig's method closes a gap in existing theory. However, LSQR is more economical on regularized problems and appears to be more reliable if the residual is not small.

In passing, we analyze a scaled “augmented system” associated with regularized problems. A bound on the condition number suggests a promising direct method for sparse equations and least-squares problems, based on indefiniteLDL T factors of the augmented matrix.

Key words

Conjugate-gradient method least squares regularization Lanczos process Golub-Kahan bidiagonalization augmented systems 

References

  1. 1.
    M. Arioli, I. S. Duff and P. P. M. de Rijk,On the augmented system approach to sparse least-squares problems, Numer. Math., 55 (1989), pp. 667–684.Google Scholar
  2. 2.
    Å. Björck,Iterative refinement of linear least squares solutions I, BIT, 7 (1967), pp. 257–278.Google Scholar
  3. 3.
    Å. Björck,A bidiagonalization algorithm for solving ill-posed systems of linear equations, Report LITH-MAT-R-80-33, Dept. of Mathematics, Linköping University, Linköping, Sweden, 1980.Google Scholar
  4. 4.
    Å. Björck,Pivoting and stability in the augmented system method, in D. F. Griffiths and G. A. Watson (eds.),Numerical Analysis 1991: Proceedings of the 14th Dundee Conference, Pitman Research Notes in Mathematics 260, Longman Scientific and Technical, Harlow, Essex, 1992.Google Scholar
  5. 5.
    J. E. Craig,The N-step iteration procedures, J. Math. and Phys., 34, 1 (1955), pp. 64–73.Google Scholar
  6. 6.
    A. Dax,On row relaxation methods for large constrained least-squares problems, SIAM J. Sci. Comp., 14 (1993), pp. 570–584.Google Scholar
  7. 7.
    I. S. Duff and J. K. Reid,Exploiting zeros on the diagonal in the direct solution of indefinite sparse symmetric linear systems, ACM Trans. Math. Softw., to appear.Google Scholar
  8. 8.
    D. K. Faddeev and V. N. Faddeeva,Computational Methods of Linear Algebra, Freeman, London, 1963.Google Scholar
  9. 9.
    R. W. Freund,Über einige CG-ähnliche Verfahren zur Lösung linearer Gleichungssysteme, Ph.D. Thesis, Universität Würzburg, FRG, 1983.Google Scholar
  10. 10.
    P. E. Gill, M. A. Saunders and J. R. Shinnerl,On the stability of Cholesky factorization for quasi-definite systems, SIAM J. Mat. Anal., 17(1) (1996), to appear.Google Scholar
  11. 11.
    G. H. Golub and W. Kahan,Calculating the singular values and pseudoinverse of a matrix, SIAM J. Numer. Anal., 2 (1965), pp. 205–224.Google Scholar
  12. 12.
    G. H. Golub and C. F. Van Loan,Unsymmetric positive definite linear systems, Linear Alg. and its Appl., 28 (1979), pp. 85–98.Google Scholar
  13. 13.
    P. C. Hansen,Test matrices for regularization methods, SIAM J. Sci. Comput., 16(2) (1995), pp. 506–512.Google Scholar
  14. 14.
    G. T. Herman, A. Lent and H. Hurwitz,A storage-efficient algorithm for finding the regularized solution of a large, inconsistent system of equations, J. Inst. Math. Appl., 25 (1980), pp. 361–366.Google Scholar
  15. 15.
    D. P. O'Leary, Private communication, 1990.Google Scholar
  16. 16.
    C. Lanczos,An iteration method for the solution of the eigenvalue problem of linear differential and integral operators, J. Res. Nat. Bur. Standards, 45 (1950), pp. 255–282.Google Scholar
  17. 17.
    P. Matstoms,Sparse QR factorization in MATLAB, ACM Trans. Math. Software, 20(1) (1994), pp. 136–159.Google Scholar
  18. 18.
    C. C. Paige,Bidiagonalization of matrices and solution of linear equations, SIAM J. Numer. Anal., 11 (1974), pp. 197–209.Google Scholar
  19. 19.
    C. C. Paige,Krylov subspace processes, Krylov subspace methods and iteration polynomials, in J. D. Brown, M. T. Chu, D. C. Ellison, and R. J. Plemmons, eds.,Proceedings of the Cornelius Lanczos International Centenary Conference, Raleigh, NC, Dec. 1993, SIAM, Philadelphia, 1994, pp. 83–92.Google Scholar
  20. 20.
    C. C. Paige and M. A. Saunders,Solution of sparse indefinite systems of linear equations, SIAM J. Numer. Anal., 12(4) (1975), pp. 617–629.Google Scholar
  21. 21.
    C. C. Paige and M. A. Saunders,LSQR: An algorithm for sparse linear equations and sparse least squares, ACM Trans. Math. Software, 8(1) (1982), pp. 43–71.Google Scholar
  22. 22.
    C. C. Paige and M. A. Saunders,Algorithm 583. LSQR: Sparse linear equations and least squares problems, ACM Trans. Math. Software, 8(2) (1982), pp. 195–209.Google Scholar
  23. 23.
    M. A. Saunders,Cholesky-based methods for sparse least squares: The benefits of regularization, Report SOL 95-1, Dept. of Operations Research, Stanford University, California, USA, 1995.Google Scholar
  24. 24.
    R. J. Vanderbei,Symmetric quasi-definite matrices, SIAM J. Optim., 5(1) (1995), pp. 100–113.Google Scholar

Copyright information

© BIT Foundation 1995

Authors and Affiliations

  • Michael A. Saunders
    • 1
  1. 1.Systems Optimization Laboratory, Department of Operations ResearchStanford UniversityStanfordUSA

Personalised recommendations