An integrated joint optimization procedure for robust and tolerance design

Many manufacturers have discovered that optimizing design parameters is a c ost-effective means of improving product quality and being competitive in t he world market. In this regard, the issues of robust design (RD) and toler ance design (TD) are clearly important, but then is significant room for im provement. The primary objective of this payer is to propose a set of enhan ced optimization strategies by combining RD and TD. To be more specific, fi rst, we consider an alternative experimental scheme using response surface methodology, while avoiding the use of controversial tools for RD such as o rthogonal arrays and signal-to-noise ratios. Secondly, we discuss an enhanc ed optimization model by simultaneously considering both the process mean a nd variance, and then show that this model provides a better (or at least e qual) solution in terms of the control factor settings. Thirdly, we show ho w the response functions for the process mean and variance, which are estim ated by using an RD principle, are transmitted into the TD stage. Fourthly, we propose an optimization model for TD and present closed-form solutions for optimum tolerance limits. Finally, we study the possible effects of maj or cost components, and observe the behaviour of the optimum control parame ter settings and the tolerance limits by carrying out sensitivity analysis.