The Convergence of a Nonmonotone Conic Trust Region Method

Abstract

In this paper, we propose a nonmonotone algorithm for unconstrained optimization that employs both conic model and trust region method. Unlike the traditional trust region method, the subproblem of our method is the conic minimization subproblem, instead of the quadratic model. The global and superlinear convergence properties of the algorithm are established under reasonable conditions.