Nonconvex notions of regularity and convergence of fundamental algorithms for feasibility problems
Abstract
We consider projection algorithms for solving (nonconvex) feasibility problems in Euclidean spaces. Of special interest are the Method of Alternating Projections (MAP) and the Douglas-Rachford or Averaged Alternating Reflection Algorithm (AAR). In the case of convex feasibility, firm nonexpansiveness of projection mappings is a global property that yields global convergence of MAP and for consistent problems AAR. Based on (\epsilon, \delta)-regularity of sets developed by Bauschke, Luke, Phan and Wang in 2012, a relaxed local version of firm nonexpansiveness with respect to the intersection is introduced for consistent feasibility problems. Together with a coercivity condition that relates to the regularity of the intersection, this yields local linear convergence of MAP for a wide class of nonconvex problems,
Keywords
Cite
@article{arxiv.1212.3349,
title = {Nonconvex notions of regularity and convergence of fundamental algorithms for feasibility problems},
author = {Robert Hesse and D. Russell Luke},
journal= {arXiv preprint arXiv:1212.3349},
year = {2014}
}
Comments
22 pages, no figures, 30 references