Analysis Of Algolrithms For Solving Unconstrained Optimization Problems:- Ijiomah, Okechukwu F.

ABSTRACT

Line search and trust region methods are known traditional methods for solving unconstrained optimization problems. In this thesis, the different algorithms for solving unconstrained optimization problems were analyzed. Also, we introduced a modified Polak-Ribiere-Polyak Conjugate gradient algorithm which makes use of both function value and gradient value information. The Armijo, Strong Wolfe and Goldstein rules problems. The numerical results obtained shows the efficiency of modified PolakRibiere-Polyak conjugate gradient method when compared with the normal PolakRibiere-Polyak conjugate gradient were applied on the line search algorithms and used to solve some standard test