A convex optimization approach for solving large scale linear systems
No Thumbnail Available
Date
2017-01
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Bulletin of Computational Applied Mathematics
Abstract
The well-known Conjugate Gradient (CG) method minimizes a strictly convex
quadratic function for solving large-scale linear system of equations when the co- efficient matrix is symmetric and positive definite. In this work we present and analyze a non-quadratic convex function for solving any large-scale linear system of equations regardless of the characteristics of the coefficient matrix. For finding the global minimizers, of this new convex function, any low-cost iterative opti- mization technique could be applied. In particular, we propose to use the low-cost globally convergent Spectral Projected Gradient (SPG) method, which allow us to extend this optimization approach for solving consistent square and rectangular linear system, as well as linear feasibility problem, with and without convex con- straints and with and without preconditioning strategies. Our numerical results indicate that the new scheme outperforms state-of-the-art iterative techniques for solving linear systems when the symmetric part of the coefficient matrix is indefi-nite, and also for solving linear feasibility problems.
Description
Keywords
Citation
Bulletin of Computational Applied Mathematics Vol.5, No.1, pp.51-74, 2017