Global Convergence Of A Spectral Hs Conjugate Gradient Method

Abstract

A spectral HS conjugate gradient method which combines the advantages of the spectral conjugate gradient method and the well-known HS conjugate gradient method has been proposed to solve general unconstrained optimization problems. It is important that the proposed method produce sufficient descent search direction at every iteration with the strong Wolfe line searches, and the global convergence for general non-convex functions can be guaranteed. Numerical results show that the proposed method is efficient and stationary by comparing with the CG-DESCENT method and the HS method, so it can be widely used in scientific computation.