Abstract:

In most engineering problems, model uncertainty is inevitably involved in the robust parameter design. Ensemble of surrogates based on the encompassing test (ET-EOS) is proposed to consider model uncertainty for response surface modeling. Firstly, sub-surrogates are assured according to the practical problem and characteristics of models, then different surrogates are constructed. Secondly, encompassing tests are used to eliminate the redundant information among surrogates and reduce the number of surrogates contained in the ensemble of surrogates, and then the effective sub-surrogates are identified. Weighted average for all models is carried out to obtain a robust ensemble model. Finally, the effectiveness of the proposed method is verified through a practical industrial example combined with a simulation example. The results reveal that the proposed method not only improves the prediction and the robustness of model prediction, but also reduce the computing cost for constructing models.