A new trust region filter algorithm (Q2378945)

A trust region filter algorithm is the subject of this paper, where the authors propose a new strategy to choose the trust region during the optimization. The choice of the trust region radius is based on previous iterates and the current gradient norm. The global convergence of the algorithm is shown under usual assumptions. Numerical results are given for nine test problems with up to 500 variables out of the CUTE collection without specifying exactly the considered problems.