RUS  ENG
Full version
JOURNALS // Proceedings of the Yerevan State University, series Physical and Mathematical Sciences // Archive

Proceedings of the YSU, Physical and Mathematical Sciences, 2014 Issue 1, Pages 7–15 (Mi uzeru43)

Mathematics

On the two-level preconditioning in least squares method

Yu. R. Akopian, R. Z. Hovhannisyan

Yerevan State University

Abstract: In the present paper an approach to construct algebraic two-level preconditioners for the matrices of normal systems arising in data fitting by least squares method with piecewise linear basis functions is proposed. The approach is based on using hierarchical grids with their subdivision into substructures and corresponding partition of the matrices. Estimates for condition numbers of preconditioned matrices are obtained.

Keywords: least squares method, normal system, condition number.

MSC: Primary 65D10; Secondary 65F08

Received: 24.09.2013
Revised: 16.10.2013

Language: English



© Steklov Math. Inst. of RAS, 2024