Non-monotone adaptive filter conic trust region method for unconstrained optimization

Abstract

In this paper, a new non-monotone trust region method based on new conic modelis proposed for unconstrained optimization, we adjust thetrust region radius by the method proposed by Hei (2003). We use filter technique to increase the chance ofaccepting the trial point. Global convergence and$Q$-superlinear convergence of the method are established. Numerical resultsreported to show the efficiency of the new method.