Preconditioned GMRES methods for least squares problems
The authors deal with the linear least squares problem: \[ \min \|\mathbf b-\mathbf A\mathbf x\|_{2}, \quad x\in \mathbb R^n,\;\mathbf A\in \mathbb R^{m\times n},\;m>n, \] where \(\mathbf A\) is a large sparse matrix with full column rank. In the first part conventional solution methods are assessed and their shortcomings ascertained. They are referenced in particular: a direct method using \(QR\) decomposition, an incomplete \(QR\) decomposition, the conjugate gradient \(CGLS\) method and the \(CR-LS(k)\) method. In the next part two versions of the generalized minimal residual (GMRES(k)) method are developed together with all proofs needed. They are based on a \(\mathbf B\in\mathbb R^{n\times m}\) mapping matrix to create a Krylov subspace (i) in the (larger) \(m\)-dimensional space and (ii) in the (smaller) \(n\)-dimensional space. Some possibilities of the matrix \(\mathbf B\) construction are discussed. As the most effective method for the matrix \(\mathbf B\) determination reveals the incomplete \(QR\) decomposition \(IMGS(l)\). Wide numerical experiments are done which enables the authors to obtain an overview concerning detailed properties of the preconditioners introduced. In the final part a relation of the parameter \(l\), preconditioning time, number of iterations and iteration time are evaluated for four \(1000\times 320, 4.9\%\) sparse matrices and seven \(10000\times 1000, 1.5\%\) sparse matrices. Results and efficiency of the proposed algorithms are presented in a form of tables, plots and very detailed commentary. It seems that the case \(l=0\) provides the best results and converges faster than conventional methods, especially when the problem is ill-conditioned. The list of references is large enough concentrating to the topic itself without any side steps to general resources.
- GMRES methods for least squares problems
- Preconditioned GMRES methods with incomplete Givens orthogonalization method for large sparse least-squares problems
- Preconditioners for least squares problems by LU factorization
- Preconditioned iterative methods for solving linear least squares problems
- Greville's method for preconditioning least squares problems
- A necessary and sufficient convergence condition of orthomin(k) methods for least squares problem with weight
- A Robust Preconditioner with Low Memory Requirements for Large Sparse Least Squares Problems
- GMRES On (Nearly) Singular Systems
- GMRES-type methods for inconsistent systems
- GMRES: A Generalized Minimal Residual Algorithm for Solving Nonsymmetric Linear Systems
- scientific article; zbMATH DE number 40775 (Why is no real title available?)
- scientific article; zbMATH DE number 852536 (Why is no real title available?)
- Incomplete Methods for Solving A^T Ax = b
- Methods of conjugate gradients for solving linear systems
- Preconditioning techniques for nonsymmetric and indefinite linear systems
- Preconditioned GMRES methods with incomplete Givens orthogonalization method for large sparse least-squares problems
- The preconditioned GMRES method for systems of coupled FEM-BEM equations
- Preconditioners for least squares problems by LU factorization
- A preconditioning strategy for the solution of linear boundary element systems using the GMRES method
- Accuracy of preconditioned CG-type methods for least squares problems.
- A Schur complement approach to preconditioning sparse linear least-squares problems with some dense rows
- Convergence and preconditioning of inexact inverse subspace iteration for generalized eigenvalue problems
- A note on preconditioned GMRES for solving singular linear systems
- Incomplete hyperbolic Gram-Schmidt-based preconditioners for the solution of large indefinite least squares problems
- Convergence analysis of GMRES methods for least squares problems
- Steepest descent preconditioning for nonlinear GMRES optimization
- GMRES methods for least squares problems
- Polynomial Preconditioned GMRES and GMRES-DR
- Preconditioning linear least-squares problems by identifying a basis matrix
- On vectorizing the preconditioned generalized conjugate residual methods
- A direct solver for the least‐squares problem arising from GMRES(k)
- scientific article; zbMATH DE number 1406963 (Why is no real title available?)
- Least squares preconditioning for mixed methods with nonconforming trial spaces
- Convergence of inner-iteration GMRES methods for rank-deficient least squares problems
- A Flexible Inner-Outer Preconditioned GMRES Algorithm
- Three-precision GMRES-based iterative refinement for least squares problems
- KRYLOV SUBSPACE METHODS WITH DEFLATION AND BALANCING PRECONDITIONERS FOR LEAST SQUARES PROBLEMS
- Preconditioned iterative methods for solving linear least squares problems
- Greville's method for preconditioning least squares problems
- New IMGS-based preconditioners for least squares problems
- Least squares and the not-normal equations
This page was built for publication: Preconditioned GMRES methods for least squares problems
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q946857)
