A practical penalty trust-region method for equality-constrained optimization problems

摘要

The ill-conditioned subproblems appeared in constrained optimization can often impede the preformance of optimization algorithms. In this paper we introduce a class of efficient iterative semi-refinement technique to solve the ill-conditioned trust-region subproblems. The nonmonotone technique is also used to increase the efficiency of the algorithm. The numerical results show that the presented method is competitive.