期刊
NONLINEAR ANALYSIS-THEORY METHODS & APPLICATIONS
卷 81, 期 -, 页码 1-11出版社
PERGAMON-ELSEVIER SCIENCE LTD
DOI: 10.1016/j.na.2012.12.009
关键词
Robust optimization; Nonsmooth Newton methods; Ellipsoidal uncertainty; Best approximations; Duality
资金
- Australian Research Council [DP-1292508]
- USA National Science Foundation [DMS-1007132]
In this paper we present a duality approach for finding a robust best approximation from a set involving interpolation constraints and uncertain inequality constraints in a Hilbert space that is immunized against the data uncertainty using a nonsmooth Newton method. Following the framework of robust optimization, we assume that the input data of the inequality constraints are not known exactly while they belong to an ellipsoidal data uncertainty set. We first show that finding a robust best approximation is equivalent to solving a second-order cone complementarity problem by establishing a strong duality theorem under a strict feasibility condition. We then examine a nonsmooth version of Newton's method and present their convergence analysis in terms of the metric regularity condition. (C) 2012 Elsevier Ltd. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据