Abstract: As the traditional comprehensive desirability model considers less of niose factors in the study for multi-responses, noise factors have been introduced into the traditional comprehensive desirability model in the paper to improve the previous one for analyging robust optimization. Firstly, response surface models of the multiple quality characteristics including controlled factors and noised factors have been conducted; secondly, based on the traditional comprehensive desirability model, we integrate multiple quality characteristics of response surface models with the improved desirability model; finally, S/N ratio has been used as an indicator of measuring robustness and optimization for the improved desirability model. The empirical study shows that the improved desirability model can be effectively conducted and can find ideal parameter combination with the help of S/N ratio.

Key words: quality management, the desirability model adding the noise factors, RSM, the desirability function, S/N ratio analysis, multi-responses, robust optimization

摘要： 针对传统的综合满意度模型在研究多响应时较少考虑噪声因素的问题,本文通过引入噪声因素对传统的满意度模型实施改进,进行多响应稳健优化分析。首先,构建包含可控因素和噪声因素的多质量特性的响应曲面模型;其次,借助传统综合满意度模型的构建方法,将多质量特性的响应曲面模型整合为改进的满意度模型;最后,将信噪比作为衡量改进的满意度模型稳健优化的指标,得到稳健优化的参数组合。实证研究表明:构建同时包含可控因素和噪声因素的改进的满意度模型是可行的,在此模型的基础上利用信噪比能够有效地找到多响应稳健优化的参数组合。

关键词: 质量管理 , 包含噪声因素的满意度模型, 响应曲面法, 满意度函数法, 信噪比分析, 多响应, 稳健优化