Black box simulation optimization : Generalized response surface methodology

The thesis consists of three papers in the area of Response Surface Methodology (RSM). The first paper deals with optimization problems with a single random objective. The contributions of that paper are a scale independent search direction and possible solutions for the step size. The second paper considers optimization problems with a single random objective and multiple random constraints, as well as deterministic box constraints. That paper extends RSM to cope with constraints, using ideas from interior point methods. Furthermore, it provides a heuristic to reach quickly a neighborhood of the optimum. The third paper copes with optimization problems with a single random objective and multiple random constraints. The contribution of that paper is an asymptotic stopping rule that tests the first-order necessary optimality conditions at a feasible point.