不同模型拟合优度下的多响应优化方法

响应曲面方法是对产品质量改进的一种有效的试验设计方法。目前,在响应模型的不同拟合优度问题基础上的多响应优化方法还有待进一步的研究。满意度函数方法是一种简单易行的响应曲面方法,而传统的满意度函数方法并未涉及响应模型的拟合优度。文章在指数满意度函数基础上提出了一种改进的满意度函数方法,以解决试验设计中不同拟合优度下的多响应优化问题。该方法将调整的Ra2dj作为响应曲面模型拟合优度的衡量标准,考虑了响应模型的拟合优度因素,分析了不同的拟合优度对多响应优化结果的影响。最后,通过实例对该方法进行验证,得到了较好的效果。

Optimization for Multiple Responses Based on Goodness of Fit for Response Surface Model

Response Surface Methodology is an efficient way of Design Of Experiment for quality improvement.More research based on response models’ goodness of fit needs to do.Desirability function methodology is a good way in RSM.However,traditional desirability function didn’t consider goodness of fit of a response model.In this paper,a modified exponential desirability function is proposed.The method considers R■■ as goodness of fit of response models and analyzes the optimization effects caused by different goodness of fit.Finally,an example is solved by the proposed method.