A breakdown of the block CG method

DOI10.1080/10556789608805643MaRDI QIDQ4354080zbMATH OpenOpenAlexFDO

Authors C. G. Broyden

Publication date 20 March 2000

Published in Optimization Methods \& Software (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1080/10556789608805643

zbMATH Keywords

block conjugate gradient methodbreakdowncomparison of methodsHestenes-Stiefel methodindefinite symmetric or nonsymmetric systemsLanczos methodstability analysis

Mathematics Subject Classification ID

Conditioning of matrices (15A12) Iterative numerical methods for linear systems (65F10) Numerical computation of matrix norms, conditioning, scaling (65F35)

Review text

The paper deals with a stability analysis of the block conjugate gradient (cg) method when applied to indefinite symmetric or nonsymmetric systems. In particular, necessary and sufficient conditions for the avoidance of a breakdown are derived. Various special cases are examined by means of these conditions and it is found that the Lanczos versions are inherently more stable than the versions due to Hestenes and Stiefel.

Recommendations

• scientific article; zbMATH DE number 1163759
• A breakdown-free block conjugate gradient method
• Avoiding breakdown in the CGS algorithm
• A stabilization of the CGS method by avoiding near-breakdown

Cites work

• A breakdown-free Lanczos type algorithm for solving linear systems
• A Completed Theory of the Unsymmetric Lanczos Process and Related Algorithms, Part I
• A Completed Theory of the Unsymmetric Lanczos Process and Related Algorithms. Part II
• A conjugate direction algorithm without line searches
• Avoiding breakdown and near-breakdown in Lanczos type algorithms
• Error Propagation in Numerical Processes
• Generating conjugate directions for arbitrary matrices by matrix equations. I
• Hyperbolic Pairs in the Method of Conjugate Gradients
• LSQR: An Algorithm for Sparse Linear Equations and Sparse Least Squares
• Matrix Analysis
• QMR: A quasi-minimal residual method for non-Hermitian linear systems
• Solution of Sparse Indefinite Systems of Linear Equations
• The block conjugate gradient algorithm and related methods
• Variable Block CG Algorithms for Solving Large Sparse Symmetric Positive Definite Linear Systems on Parallel Computers, I: General Iterative Scheme

Cited in

(5)

• An analysis of the composite step biconjugate gradient method
• A review of formal orthogonality in Lanczos-based methods
• A breakdown-free block conjugate gradient method
• Enhancing Block Cimmino for Sparse Linear Systems with Dense Columns via Schur Complement
• Block CG algorithms revisited

This page was built for publication: A breakdown of the block CG method

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4354080)