[Home]   [down2  News]   [down2  Events]   [down2  People]   [down2  Research]   [down2  Education]   [Visitor Info]   [UCSD Only]   [Admin]
Home > Events > CCoM > Abstract

Flexible Penalty Methods for Nonlinear Optimization

Slava Kungertsev
UCSD

Abstract:

Penalty functions and filter methods are two commonly-used algorithms for solving nonlinear constrained problems. In practice they suffer from a number of disadvantages that slow their convergence. Flexible penalty functions are an attempt to utilize the best characteristics of both and have been demonstrated to exhibit good global convergence properties.

Tuesday, January 12, 2010
11:00AM AP&M 2402