Stochastic Kriging for Simulation Metamodeling

摘要: We extend the basic theory of kriging, as applied to design and analysis deterministic computer experiments, stochastic simulation setting. Our goal is provide flexible, interpolation-based metamodels output performance measures functions controllable or decision variables, uncontrollable environmental variables. To accomplish this, we characterize both intrinsic uncertainty inherent in a extrinsic about unknown response surface. use tractable examples demonstrate why it critical types uncertainty, derive general results for experiment analysis, present numerical example that illustrates kriging method.