Strong global convergence of an adaptive nonmonotone memory gradient method

摘要

In this paper, we develop an adaptive nonmonotone memory gradient method for unconstrained optimization. The novelty of this method is that the stepsize can be adjusted according to the characteristics of the objective function. We show the strong global convergence of the proposed method without requiring Lipschitz continuous of the gradient. Our numerical experiments indicate the method is very encouraging. (c) 2006 Elsevier Inc. All rights reserved.