Metric regularity relative to a cone (Q2330981)

Metric regularity and related properties are powerful tools for dealing with problems of optimization and variational analysis, and its has a long history. Important applications enclose the study of stability of variational systems as well as convergence of Newton's type methods. In recent years, due to algorithmic purposes, an increasing attention has been paid to metric regularity concepts, and, for instance, some kinds of relative metric regularity to quantitative convergence analysis of algorithms are developed. In this work, the authors consider the particular case of relative metric regularity called metric regularity relative to a cone, that is a natural extension of directional metric regularity by replacing a direction by a cone. After some basic definitions and tools, a slope criterion for metric regularity relative to a set is given. Then the authors prove that metric regularity relative to a cone is stable under perturbation by a Lipschitz continuous function, and they give a coderivative characterization of metric regularity relative to a cone.