A new hybrid CG method as convex combination

Abstract

Conjugate gradient methods are among the most efficient methods for solving optimization models. In this paper, a newly proposed conjugate gradient method is proposed for solving optimization problems as a convex combination of the Harger-Zhan and Dai-Yaun nonlinear conjugate gradient methods, which is capable of producing the sufficient descent condition with global convergence properties under the strong Wolfe conditions. The numerical results demonstrate the efficiency of the proposed method with some benchmark problems.