On the Convergence Behavior of IDR(s) and Related Methods

An explanation is given of the convergence behavior of IDR(s) methods. The convergence mechanism of these algorithms has two components. The first consists of damping properties of certain factors in the residual polynomials, which becomes less important for large values of s. The second component depends on the behavior of Lanczos polynomials that occur in the theoretical description. This part of the residual polynomials is related to Lanczos methods with s left starting vectors, as described in a paper by Yeung and Chan on their ML(k)BiCGSTAB method, in [SIAM J. Sci. Comput., 21 (1999), pp. 1263--1290]. In this paper, the behavior of the second component is compared with the full GMRES method [SIAM J. Sci. Statist. Comput., 7 (1986), pp. 856--869] and an expected rate of convergence is given, based on a random choice of the s shadow vectors

[1]  Gerard L. G. Sleijpen,et al.  Maintaining convergence properties of BiCGstab methods in finite precision arithmetic , 1995, Numerical Algorithms.

[2]  Peter Sonneveld,et al.  On the statistical properties of solutions of completely random linear systems , 2010 .

[3]  G. Golub,et al.  Gmres: a Generalized Minimum Residual Algorithm for Solving , 2022 .

[4]  Homer F. Walker,et al.  Residual smoothing and peak/plateau behavior in Krylov subspace methods , 1995 .

[5]  Y. Saad,et al.  GMRES: a generalized minimal residual algorithm for solving nonsymmetric linear systems , 1986 .

[6]  Qiang Ye,et al.  Residual Replacement Strategies for Krylov Subspace Iterative Methods for the Convergence of True Residuals , 2000, SIAM J. Sci. Comput..

[7]  Tijmen P. Collignon,et al.  Minimizing synchronization in IDR (s) , 2011, Numer. Linear Algebra Appl..

[8]  Gerard L. G. Sleijpen,et al.  BiCGstab(l) and other hybrid Bi-CG methods , 1994, Numerical Algorithms.

[9]  M. Eiermann,et al.  Geometric aspects of the theory of Krylov subspace methods , 2001, Acta Numerica.

[10]  Tony F. Chan,et al.  ML(k)BiCGSTAB: A BiCGSTAB Variant Based on Multiple Lanczos Starting Vectors , 1999, SIAM J. Sci. Comput..

[11]  Gerard L. G. Sleijpen,et al.  Bi-CGSTAB as an induced dimension reduction method , 2010 .

[12]  H. V. der.  Residual Replacement Strategies for Krylov Subspace Iterative Methods for the Convergence of True Residuals , 2000 .

[13]  Martin B. van Gijzen,et al.  IDR(s): A Family of Simple and Fast Algorithms for Solving Large Nonsymmetric Systems of Linear Equations , 2008, SIAM J. Sci. Comput..

[14]  Nathan Halko,et al.  Finding Structure with Randomness: Probabilistic Algorithms for Constructing Approximate Matrix Decompositions , 2009, SIAM Rev..

[15]  Gerard L. G. Sleijpen,et al.  Flexible and multi-shift induced dimension reduction algorithms for solving large sparse linear systems , 2011 .

[16]  A. Greenbaum.  Estimating the Attainable Accuracy of Recursively Computed Residual Methods , 1997, SIAM J. Matrix Anal. Appl..

[17]  J. Cullum,et al.  A block Lanczos algorithm for computing the q algebraically largest eigenvalues and a corresponding eigenspace of large, sparse, real symmetric matrices , 1974, CDC 1974.

[18]  Peter N. Brown,et al.  A Theoretical Comparison of the Arnoldi and GMRES Algorithms , 1991, SIAM J. Sci. Comput..

[19]  Henk A. van der Vorst,et al.  Transpose-free formulations of Lanczos-type methods for nonsymmetric linear systems , 2004, Numerical Algorithms.

[20]  Anne Greenbaum,et al.  Relations between Galerkin and Norm-Minimizing Iterative Methods for Solving Linear Systems , 1996, SIAM J. Matrix Anal. Appl..

[21]  Marlis Hochbruck,et al.  Error Analysis of Krylov Methods In a Nutshell , 1998, SIAM J. Sci. Comput..

[22]  Gene H. Golub,et al.  The block Lanczos method for computing eigenvalues , 2007, Milestones in Matrix Computation.

[23]  Martin H. Gutknecht,et al.  Eigenvalue Computations Based on IDR , 2013, SIAM J. Matrix Anal. Appl..

[24]  Masaaki Sugihara,et al.  GBi-CGSTAB(s, L): IDR(s) with higher-order stabilization polynomials , 2010, J. Comput. Appl. Math..

[25]  P. Wesseling,et al.  Numerical experiments with a multiple grid and a preconditioned Lanczos type method , 1980 .

[26]  Martin H. Gutknecht,et al.  Variants of BICGSTAB for Matrices with Complex Spectrum , 1993, SIAM J. Sci. Comput..

[27]  Valeria Simoncini,et al.  Interpreting IDR as a Petrov--Galerkin Method , 2010, SIAM J. Sci. Comput..

[28]  Malay Ghosh,et al.  A Simple Derivation of the Wishart Distribution , 2002 .

[29]  Gerard L. G. Sleijpen,et al.  Exploiting BiCGstab(ℓ) Strategies to Induce Dimension Reduction , 2010, SIAM J. Sci. Comput..

[30]  Martin H. Gutknecht,et al.  Look-Ahead Procedures for Lanczos-Type Product Methods Based on Three-Term Lanczos Recurrences , 2000, SIAM J. Matrix Anal. Appl..

[31]  A. Edelman.  Eigenvalues and condition numbers of random matrices , 1988 .

[32]  Henk A. van der Vorst,et al.  Bi-CGSTAB: A Fast and Smoothly Converging Variant of Bi-CG for the Solution of Nonsymmetric Linear Systems , 1992, SIAM J. Sci. Comput..

[33]  Gerard L. G. Sleijpen,et al.  Reliable updated residuals in hybrid Bi-CG methods , 1996, Computing.

[34]  D. R. Fokkema,et al.  BiCGstab(ell) for Linear Equations involving Unsymmetric Matrices with Complex Spectrum , 1993 .

[35]  Martin B. van Gijzen,et al.  Algorithm 913: An elegant IDR(s) variant that efficiently exploits biorthogonality properties , 2011, TOMS.