关键词: 贝叶斯模型平均, 因子效应原则, 模型不确定性, 稳健参数设计

Abstract: Aiming at the robust parameter design problem under model uncertainty,on the basis of Bayesian model averaging method,Bayesian Model Averaging based on Effect Principle (BMA-EP) robust design methodology was proposed by taking the factor effect principle into account.With prior information and Bayesian law,the main effects'posterior probabilities and a general BMA model were obtained.According to these posterior probabilities,all main effects'prior probabilities were updated by using effect hierarchy principle and effect heredity principle step by step,and the model's posterior probabilities were identified.With these posterior probabilities as weights,weighted average for all models was carried out to obtain the robust model which had better predicted performance and experiment design principle.The effectiveness of proposed method was verified through a practical industrial example combined with a simulation example.The results revealed that the method not only improved the performance of model prediction,but also increased the reliability of optimal parameter settings.

Key words: Bayesian model averaging, factor effect principle, model uncertainty, robust parameter design