基于均方根误差建模的多响应稳健参数设计

针对复杂产品质量设计阶段的静态多响应稳健参数设计中的模型参数不确定性问题,现有的多响应优化方法大都对响应的样本均值与样本方差采用双响应曲面法分别建模以考虑多响应的最优性与稳健性。文章在此基础上分析构建了均方根误差响应这一新的稳健性度量指标,提出了考虑模型参数不确定性的满意度函数方法,结合置信区间思想分析了模型参数不确定性对均方根误差响应的影响,并依据复杂产品质量设计阶段的实例进行分析研究,验证了该方法能够得到多响应系统在模型参数不确定性情况下更为稳健的全局最优解。

Multi-response Robust Parameter Design Based on RMS Error Modeling

Most of the existing multi-response optimization methods use the two-response surface method to model the sample mean and sample variance of the response respectively to consider the optimality and robustness of the multi-response.On this basis,and in order to solve the problem of model parameters uncertainty in the static multi-response robust parameter design of complex product quality design stage,this paper analyzes and constructs a new robustness measurement index of root mean square(RMS)error response,proposes a satisfaction function method considering the uncertainty of model parameters,and also combines with the idea of confidence interval to analyze the influence of the uncertainty of model parameters on RMS error response.Finally the paper conducts an analysis and research based on the example of complex product quality design stage,which verifies that the proposed method can obtain a more robust global optimal solution for multi-response systems under the uncertainty of model parameters.