A Modified NPRP Nonlinear Conjugate Gradient Method with Global Convergence Properties

Abstract

Conjugate gradient method (CG), is one of many methods used to solve large scale optimization problems in many fields. In fact memory storage and speed of the algorithm to reach the optimum solution for general nonlinear unconstrained optimization functions are our goals. In this paper, we modified a new efficient method passes the sufficient decent condition, and it has global convergence under exact line search and strong Wolfe line search. Furthermore the numerical results to find the optimum solution of some test functions show the new modification has the best results in CPU time, and the number of iteration when it comparing with other conventional methods.