Two-stage three-way enhanced multi-criteria classification optimization for risk-averse product design programming

Abstract

Product design is essential in the product lifecycle. In product design, decision errors from designers are ineluctable due to the lack of sufficient knowledge and experience on design parameters. However, the penalties (e.g., expected risk loss costs) of design decisions have been overlooked as always. The result is that being risk-averse naturally becomes a major issue to be advocated. Meanwhile, the preferences of designers, representing rational cognition of product design, are crucial in product design. In response to the above two critical issues, a two-stage three-way enhanced multi-criteria classification optimization for risk-averse product design programming is proposed. In the first stage, the three-way decisions (3WD) theory is introduced to handle the relation between the penalties and design decisions, and provide an uncertain decision representing an intermediate attitude of designers on design parameters. Additionally, designers are endowed to preferentially consider design parameters that are profitable for the success of product design. The purpose of the first stage is to formulate multiple alternative design schemes. In the second stage, under the framework of the ELimination Et Choix Traduisant la REalité (ELECTRE) TRI, a multi-criteria classification optimization is constructed to cater to designers’ preferences. Subsequently, we can identify the optimal design scheme from these multiple alternative design schemes. The effectiveness and merits of the proposed approach are demonstrated with two illustrative examples and a substrate design case. As shown from the case studies, the proposed approach has a stable performance in terms of identifying the optimal design scheme in a stochastic environment.