A New Nonmonotone Trust Region Algorithm for Solving Unconstrained Optimization Problems
DOI:
https://doi.org/10.4208/jcm.1401-m3975Keywords:
Unconstrained optimization problems, nonmonotone trust region method, global convergence, superlinear convergence.Abstract
Based on the nonmonotone line search technique proposed by Gu and Mo (Appl. Math. Comput. 55, (2008) pp. 2158-2172), a new nonmonotone trust region algorithm is proposed for solving unconstrained optimization problems in this paper. The new algorithm is developed by resetting the ratio $ρ_k$ for evaluating the trial step $d_k$ whenever acceptable. The global and superlinear convergence of the algorithm are proved under suitable conditions. Numerical results show that the new algorithm is effective for solving unconstrained optimization problems.