In this paper, a reduction-type method combined with a line search filter method to solve nonlinear semi-infinite programming problems is presented. The algorithm uses the simulated annealing method equipped with a function stretching technique to compute the maximizers of the constraint, and a penalty technique to solve the reduced finite optimization problem. The filter method is used as an alternative to merit functions to guarantee convergence from poor starting points. Preliminary numerical results with a problems test set are shown.
展开▼
