A proposed response surface-based robust design model for quality engineering problems

The aim of robust design models is to reduce the variability reduction as small as possible. The process bias defined as a difference between the desired target value and the process mean is an important concern for quality engineering problems. In addition, the selection of different variability measures may also change optimal operating conditions for a response variable. Therefore, this paper is three-fold. One, another view of the dual response model is proposed with the three different variability measures in order to determine optimum robust design solutions for input variables while minimizing the process bias. Two, the linearization of constraints is performed using the sequential quadratic programming method as an effective optimization method. Three, a printing process from the literature is conducted to obtain the best optimal settings for input variables. Finally, the results of the proposed model show approximately % 16 more variance reduction than traditional models.