Abstract
We propose a two-phase acceleration technique for the solution of Symmetric and Positive Definite linear systems with multiple right-hand sides. In the first phase we compute some partial spectral information related to the ill conditioned part of the given coefficient matrix and, in the second phase, we use this information to improve the convergence of the Conjugate Gradient algorithm. This approach is adequate for large scale problems, like the simulation of time dependent differential equations, where it is necessary to solve consecutively several linear systems with the same coefficient matrix (or with matrices that present very close spectral properties) but with changing right-hand sides. To compute the spectral information, in the first phase, we combine the block Conjugate Gradient algorithm with the Inexact Subspace Iteration to build a purely iterative algorithm, that we call BlockCGSI. We proceed to an inner-outer convergence analysis and we show that it is possible to determine when to stop the inner iteration in order to achieve the targeted invariance in the outer iteration. The spectral information is used in a second phase to remove the effect of the smallest eigenvalues in two different ways: either by building a Spectral Low Rank Update preconditioner, or by performing a deflation of the initial residual in order to remove part of the solution corresponding to the smallest eigenvalues.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Parlett, B.N.: The Symmetric Eigenvalue Problem. SIAM, Philadelphia (1998)
O’Leary, D.P.: The block conjugate gradient algorithm and related methods. Linear Algebra Appl. 29, 293–322 (1980)
Arioli, M., Duff, I., Ruiz, D., Sadkane, M.: Block Lanczos techniques for accelerating the block cimmino method. SIAM J. Sci. Stat. Comput. 16(6), 1478–1511 (1995)
Balsa, C., Daydé, M., Palma, J.M.L.M., Ruiz, D.: Monitoring the block conjugate gradient convergence within the inexact inverse subspace iteration. In: Wyrzykowski, R., Dongarra, J., Meyer, N., Wasniewski, J. (eds.) Parallel Processing and Applied Mathematics. Lectures Notes in Computer Science - PPAM05, LNCS 3911, pp. 494–504. Springer, Berlin (2006)
Arioli, M., Ruiz, D.: Block conjugate gradient with subspace iteration for solving linear systems. In: Margenov, S., Vassilevski, P. (eds.) Iterative Methods in Linear Algebra, Second IMACS Symposium on Iterative Metohds in Linear Algebra, pp. 64–79. Blagoevgrad, Bulgaria (1995)
Giraud, L., Ruiz, D., Touhami, A.: A comparative study of iterative solvers exploiting spectral information for SPD systems. SIAM J. Sci. Comput. 27(5), 1064–8275 (2006)
Carpentieri, B., Duff, I., Giraud, L.: A class of spectral two-level preconditioners. SIAM J. Sci. Comput. 25(2), 749–765 (2003)
Lai, Y.L., Lin, K.Y., Lin, W.W.: An inexact inverse iteration for large sparse eigenvalue problems. Numer. Linear Algebra Appl. 4(5), 425–437 (1997)
Smit, P., Paardekooper, M.H.C.: The effects of inexact solvers in algorithms for symmetric eigenvalue problems. Linear Algebra Appl. 287, 337–357 (1999)
Golub, G.H., Ye, Q.: Inexact inverse iteration for generalized eigenvalue problems. BIT 40, 671–684 (2000)
Simoncini, V., Eldén, L.: Inexact Rayleigh quotient-type methods for eigenvalue computations. BIT 42, 159–182 (2002)
Berns-Mueller, J., Graham, I. G., Spence, A.: Inexact inverse iteration for symmetric matrices. Linear Algebra and its Appl. 416(2–3), 389–413 (2006)
Notay, Y.: Combination of Jacobi-Davidson and conjugate gradients for the partial symmetric eigenproblem. Numer. Linear Algebra Appl. 9, 21–44 (2002)
Dongarra, J.J., Ducroz, J., Hammarling, S., Duff, I.: A set of level 3 basic linear algebra subprograms. ACM Trans. Math. Softw. 16(1), 1–28 (1990)
Rigal, J.L., Gaches, J.: On the compatibility of a given solution with the data of a linear system. J. ACM 14(3), 543–548 (1967)
Hageman, L.A., Young, D.M.: Applied Iterative Methods. Academic Press, New York (1981)
Van der Sluis, A., van der Vorst, H.A.: The rate of convergence of Conjugate Gradients. Numer. Math. 48(5), 543–560 (1986)
Dongarra, J.J., Ducroz, J., Hammarling, S., Hanson, R.: An extended set of Fortran basic linear algebra subprograms. ACM Trans. Math. Softw. 14, 1–17 (1988)
Balsa, C., Braza, M., Daydé, M., Palma, J.M.L.M., Ruiz, D.: Improving the numerical simulation of an airflow problem with the BlockCGSI algorithm. In: Daydé, M., Palma, José M.L.M., Coutinho, Álvaro L.G.A., Pacitti, Esther, Correia Lopes, J. (eds.) High Performance Computing for Computational Science - VECPAR 2006. LNCS 4395, pp. 281–291. Springer, Berlin (2007)
Notay, Y.: Convergence analysis of inexact Rayleigh quotient iteration. SIAM J. Matrix Anal. Appl. 24(3), 627–644 (2003)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer International Publishing AG, part of Springer Nature
About this paper
Cite this paper
Balsa, C., Daydé, M., Palma, J.M.L.M., Ruiz, D. (2018). Inexact Subspace Iteration for the Consecutive Solution of Linear Systems with Changing Right-Hand Sides. In: Pinto, A., Zilberman, D. (eds) Modeling, Dynamics, Optimization and Bioeconomics III. DGS BIOECONOMY 2016 2015. Springer Proceedings in Mathematics & Statistics, vol 224. Springer, Cham. https://doi.org/10.1007/978-3-319-74086-7_3
Download citation
DOI: https://doi.org/10.1007/978-3-319-74086-7_3
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-74085-0
Online ISBN: 978-3-319-74086-7
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)