A Modified Fletcher-Reeves Conjugate Gradient Method For Unconstrained Optimization
Conjugate gradient (CG) method is one of many tools used to solve large optimization unconstrained problems.
From 1952 until now many methods appeared to improve CPU time and number of iteration which is needed to reach the
optimum solution. In this paper we depict a new positive CG method derived from Fletcher–Reeves (FR) method, the new
coefficient achieves the global convergence properties under exact line search. In addition, it possesses the sufficient descent
condition. The numerical computations where the step size achieves by exact line search show the efficiency of the new
method is better than FR method; furthermore it solves all test functions.
Keywords— Conjugate Gradient, Global Convergence, Exact Line Search, Optimum Solution.