Automated response surface methodology for stochastic optimization models with unknown variance

Response surface methodology (RSM) is an optimization tool that was introduced in the early 50's by Box and Wilson (1951). In this paper we are interested in finding the best settings for an automated RSM procedure when there is very little information about the objective function. We present a framework of the RSM procedures that is founded in recognizing local optima in the presence of noise. We emphasize both stopping rules and restart procedures. The results show that considerable improvement is possible over the proposed settings in the existing literature.