NONMONOTONE TRUST-REGION ALGORITHMS FOR NONLINEAR OPTIMIZATION SUBJECT TO CONVEX CONSTRAINTS

This paper presents two new trust-region methods for solving nonlinear optimization problems over convex feasible domains. These methods are distinguished by the fact that they do not enforce strict monotonicit y of the objective function values at successive iterates. The algorit hms are proved to be convergent to critical points of the problem from any starting point. Extensive numerical experiments show that this ap proach is competitive with the LANCELOT package. (C) 1997 The Mathemat ical Programming Society, Inc.