05658710

j

05658710 Shen, Chungen Xue, Wenjuan Pu, Dingguo A globally convergent trust region multidimensional filter SQP algorithm for nonlinear programming. Int. J. Comput. Math. 86, No. 12, 2201-2217 (2009). 2009 Taylor \& Francis, Abingdon EN trust region multidimensional filter SQP KKT point nonlinear programming

doi:10.1080/00207160802702400

A trust-region multidimensional filter method for solving nonlinear constrained optimization problem with equation and inequality constraints is proposed. In order to obtain a better numerical performance, the constraints are partitioned in $p$ parts so that one may set different parameters for different parts of constraints correspondingly. Another important feature of the algorithm is that the entries in the filter are independent of the objective function. Besides, the algorithm uses the non-monotonic technique for accepting trial steps, which increases the flexibility of acceptance criteria. Global convergence of the proposed algorithm under mild conditions is proved. Numerical results presented in the last section of the paper confirm the robustness of the proposed approach. Karel Zimmermann (Praha)