A nonmonotone trust region method based on simple conic models for unconstrained optimization

Abstract

A new nonmonotone trust region algorithm with simple conic models for unconstrained optimization is proposed. Compared to traditional conic trust region methods, the new method needs less memory capacitance and computational complexity. The global convergence and fast local convergence rate of the proposed algorithm are established under some reasonable conditions. Numerical tests indicate that the new algorithm is efficient and robust.