The Convergence Properties for Regularized Landweber Method

The Convergence Properties for Regularized Landweber Method

0.00 Avg rating0 Votes
Article ID: iaor20163677
Volume: 171
Issue: 1
Start Page Number: 262
End Page Number: 275
Publication Date: Oct 2016
Journal: Journal of Optimization Theory and Applications
Authors:
Keywords: programming: linear, heuristics, matrices
Abstract:

Landweber scheme is a widely used method to get a stable solution of linear system. The iteration of the Landweber scheme is viewed as a solution of normal equation for a least‐squares functional. However, in practice, regularized least‐squares functional is considered so as to get a more suitable solution. In this paper, we consider a regularized optimization problem and study the regularized Landweber scheme. Using the eigenvalue decomposition and the result that two symmetric semi‐positive definite matrices can be diagonalized simultaneously, we derive a presentation of the regularized Landweber scheme and then generate the convergence properties for the regularized Landweber iteration. Finally, we apply two‐dimensional numerical examples to confirm the convergence conditions.

Reviews

Required fields are marked *. Your email address will not be published.