The global convergence properties of an improved conjugate gradient method

Abstract

Conjugate gradient (CG) methods have played a significant role in solving large scale unconstrained optimization. This is due to its simplicity, low memory requirement, and global convergence properties. Various studies and modifications have been done recently to improve this method. In this paper, we proposed a new conjugate gradient parameter ) ( k  which possesses global convergence properties under the exact line search. Numerical result shows that our new formula performs better when compared to other classical conjugate gradient methods.