An interactive preference decision making approach to multi-response process design with location and dispersion effects

Abstract

Setting of design variables to meet the desired region of quality characteristics is a common problem in quality control. In this regard, many different optimisation approaches are available in literature. In most of traditional approaches, the required preference information should be articulated by the decision maker (DM) in advance. However, pre-determining such parameters is usually difficult and inaccurate. Furthermore, the majority of these techniques assume the constant variance of responses over the experimental region which is not a real assumption. Therefore, an interactive method for optimising the multiple response problems is presented in the current paper which does not require any signification of DM's preference information before the solving process. This method considers the location and dispersion effects of quality characteristics along with specification limits in a unified framework based on desirability function and the concept of coefficient of variation. The obtained results are compared with some of the existing methods on a real example to display the efficiency of the proposed method.