A strongly convergent hybrid proximal method in Banach spaces

被引:15
|
作者
Otero, RG
Svaiter, BF
机构
[1] Univ Fed Rio de Janeiro, Inst Econ, BR-22290240 Rio De Janeiro, Brazil
[2] Inst Matematica Pura & Aplicada, BR-22460320 Rio De Janeiro, Brazil
关键词
proximal point method; relative error; inexact solutions; hybrid steps; strong convergence; enlargement of maximal monotone operators;
D O I
10.1016/j.jmaa.2003.09.030
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
This paper is devoted to the study of strong convergence in inexact proximal like methods for finding zeroes of maximal monotone operators in Banach spaces. Convergence properties of proximal point methods in Banach spaces can be summarized as follows: if the operator have zeroes then the sequence of iterates is bounded and all its weak accumulation points are solutions. Whether or not the whole sequence converges weakly to a solution and which is the relation of the weak limit with the initial iterate are key questions. We present a hybrid proximal Bregman projection method, allowing for inexact solutions of the proximal subproblems, that guarantees strong convergence of the sequence to the closest solution, in the sense of the Bregman distance, to the initial iterate. (C) 2003 Elsevier Inc. All rights reserved.
引用
收藏
页码:700 / 711
页数:12
相关论文
共 50 条