A method for stochastic constrained optimization using derivative-free surrogate pattern search and collocation (Q975138)

A new stochastic derivative-free optimization method is proposed. The method can be applied to physical problems, for which evaluating cost functions involves significant computational expense. The method extends the previously developed surrogate management framework to allow uncertainties in both simulation parameters and design variables. The concept of stochastic spaces for representing and quantifying random variables is used. The proposed technique is tested on four numerical optimization problems. It is shown that the technique has significantly better efficiency in comparison with traditional Monte Carlo schemes. Problems with multiple probabilistic constraints are also discussed.