Strong global convergence of an adaptive nonmonotone memory gradient method (Q870222)

This article presents an iterative gradient method to optimize an unconstrained problem of a continuously differentiable function. The article begins with a short introduction which outlines the necessary background information. In section~2 by the main details of a non-monotone memory gradient algorithm. In the proposed algorithm, the stepsize of the method can be adjusted accordingly to the characteristics of the objective function. The article concludes with a study of global convergence of the presented algorithm and a series of numeric tests.