QP-Based Methods for Large-Scale Nonlinearly Constrained Optimization.

Abstract
Several methods for nonlinearly constrained optimization have been suggested in recent years that are based on solving a quadratic programming (QP) subproblem to determine the direction of search. Even for dense problems, there is no consensus at present concerning the 'best' formulation of the QP subproblem. When solving large problems, many of the options possible for small problems become unreasonably expensive in terms of storage and/or arithmetic operations. This paper discusses the inherent difficulties of developing QP-based methods for large-scale nonlinearly constrained optimization, and suggests some possible approaches. (Author)

This publication has 0 references indexed in Scilit: