Global Convergence of a Class of Trust Region Algorithms for Optimization Using Inexact Projections on Convex Constraints

Abstract
. A class of trust region based algorithms is presented for the solution of nonlinearoptimization problems with a convex feasible set. At variance with previously published analysisof this type, the theory presented allows for the use of general norms. Furthermore, the proposedalgorithms do not require the explicit computation of the projected gradient, and can thereforebe adapted to cases where the projection onto the feasible domain may be expensive to calculate.Strong global convergence ...