Two uncertain chance-constrained programming models to setting target levels of design attributes in quality function deployment

Abstract

Quality function deployment (QFD) is widely acknowledged as a customer-oriented product design tool, which is generated by translating consumer demands into design attributes of a product. In order to depict the internal ambiguous factors in the development process more appropriately, uncertain variables with a specialized kind of regular uncertainty distributions based on uncertainty theory are applied. Subsequently, two uncertain chance-constrained programming (CCP) models used for formulating the QFD procedure are set forth, whose objectives are maximizing the consumer satisfaction and minimizing the design cost, respectively. To demonstrate the feasibility of the proposed modelling approach, an example of the motorcycle design problem is illustrated, in which the new target levels of design attributes are selected and analyzed according to the decision-makers’ subjectivity and preference at different confidence levels. Additionally, a comparative study between the uncertain CCP approach and another uncertain expected value modelling approach is conducted. The results indicate that uncertain CCP models are more suitable for optimization in the QFD procedure.