An Improved Search Direction Method for solving Nonlinear Program based on Generalized Reduced Gradient

Abstract

The aim of this paper is to develop an approach using feasible direction for solving a constrained nonlinear program. The developed method is carried out with the enhancements of determining the search direction and the line search steps. In particular the method combines efficient sparse-matrix approach which can be found in the revised simplex method with stable reduced-gradient methods for tackling the nonlinearities.