A New Hybrid Method for Large Scale and Ill-posed Problems

摘要：

Hybrid methods have been proved as effective approaches for solving large scale and ill-posed problems.Combing an iterative method with a direct regularization method on the projected problem, they are able to stabilize the semi-convergence behavior of iterative regularization method. In this paper, we present an efficient hybrid method, based on Lanczos bidiagonalization and modified singular value decomposition (SVD) regularization.Experimental results show that the proposed method is effective on many applications.

关键词： ill-posed problems Tikhonov regularization LSQR SVD

作者: Jianjun Zhang Qin Wang

作者单位: Department of Mathematics, Shanghai University, Shanghai 200444, China

会议类型: 国际会议

会议名称: The Third International Workshop on Applied Matriz Theory(第三届国际矩阵分析与应用会议)

会议地点: 杭州

会议语种:英文

页码: 1326-1329

在线出版日期: 2009-07-09（万方平台首次上网日期，不代表论文的发表时间）