Two new conjugate gradient methods in unconstrained optimization problems

Abstract

Zheng Y. and Zheng B. in [14] modified Dai-Liao conjugate gradient method to come up with two new Dai-Liao-type conjugate gradient methods. These methods were shown to have satisfied descent condition taken into consideration the strong Wolfe line search. Convergence for objective functions were also guarantied. In this work, two new conjugate gradient methods are introduced in line with the work of Zheng Y. and Zheng B. [14] by changing the first term in AyO-CG method [3] to solve unconstrained non-linear optimization problems. Descent properties of these methods are shown and guarantied. Convergence analyses of these methods in line with strong Wolfe conditions showed that they are globally convergent. Comparison based on Dolan More performance profile of the numerical strength of these methods with the two modified Dai-Laio type methods proved that our methods compare favorably well with them