zbMATH — the first resource for mathematics

Smoothing-norm preconditioning for regularizing minimum-residual methods. (English) Zbl 1154.65028
The paper is devoted to the development of specific preconditioners for minimum residual methods in the context of large-scale discrete ill-posed problems yielding a squared matrix. In particular, a preconditioning technique which is well established for conjugate gradient least squares methods is adapted to be suitable for generalized minimal residual (GMRES) methods and related algorithms. Numerical experiments are presented to confirm the performance of the scheme.

65F35 Numerical computation of matrix norms, conditioning, scaling
65F22 Ill-posedness and regularization problems in numerical linear algebra
65F10 Iterative numerical methods for linear systems
Full Text: DOI