Using DEA and GA Algorithm for Finding an Optimal Design Chain Partner Combination

Abstract

Abstract. To assist enterprises in building the optimal design chain partner combination, this research focuses on the development of a weight restricted DEA model, in which appropriate design chain partners are evaluated and selected according to different partner roles, and appropriate partner sets are formed correspondingly. As product development time and costs depend closely on coordination efficiency among different design chain partner members, this study takes this factor into account by developing a multi-objective design chain partner combination evaluation and selection model. This study uses multiobjective genetic algorithm (MOGA) to search for the optimal design chain partner combination in order to achieve the objective of minimized new product development costs and time in conjunction with maximized product reliability.

Partner selection model for design chain collaboration

“…Finally, a notebook computer case study was proposed to verify their model and derive managerial implications. Chuang et al (2009) also built a design partner selection model to assist enterprises in selecting the optimal partners. First, they addressed a weight-restricted DEA model, in which different criteria assessment systems were proposed based on different partner roles.…”

Classification and mitigation of uncertainty as per the product design stages: framework and case study

“…Of late, the complexity of product design has significantly increased, and product development requires significant knowledge of engineering, information technology, and design, apart from in-depth product functional knowledge. At times, it is beneficial for companies to outsource the product development activities to specialists, as per the case study given by Chiang [36]. For this he proposed multi-objective decision-making methodology to create an optimal design chain partner combination by employing a weight restricted data envelopment analysis (DEA) approach.…”

“…For multidomain systems, Agarwal et al [37] proposed an iterative solution in the form of evidence theory to quantify uncertainty due to uncertain measures of conviction and conceivability. Chiang [36] and Agarwal et al [37] indicated that multi-domain interaction uncertainty may be managed by accruing proper knowledge of the missing domain information. Hence, multi-domain interaction uncertainty may also be classified as epistemic uncertainty.…”