[Home]   [  News]   [  Events]   [  People]   [  Research]   [  Education]   [Visitor Info]   [UCSD Only]   [Admin]
Home > Research > Publications > Abstract
Search this site:


Directors:
Randolph E. Bank
Philip E. Gill
Michael Holst

Administrative Contact:
Terry Le

Office: AP&M 7431
Phone: (858)534-9813
Fax: (858)534-5273
E-mail: tele@ucsd.edu
Distance-to-Solution Estimates for Optimization Problems with Constraints in Standard Form

by Philip E. Gill, Vyacheslav Kungurtsev, Daniel P. Robinson

Abstract:

An important tool in the formulation and analysis of algorithms for constrained optimization is a quantity that provides a practical estimate of the distance to the set of primal-dual solutions. Such ``distance-to-solution estimates'' may be used to identify the inequality constraints satisfied with equality at a solution, and to formulate conditions used to terminate a sequence of solution estimates. This note concerns the properties of a particular distance-to-solution estimate for optimization problems with constraints written in so-called ``standard form'', which is a commonly-used approach for formulating constraints with a mixture of equality and inequality constraints.

UCSD-CCoM-16-01.pdf   April 2016