[KCI] The Optimal Common Attribute Selection in Product Design by Logic-Based Benders Decomposition

posted Mar 16, 2015, 5:15 AM by CWS Yonsei   [ updated Mar 16, 2015, 5:51 AM ]
In the modern product design, determining a platform that all variety of products are based on is a critical issue. This paper proposes an algorithm to solve the optimal common attribute selection in the product design by extending the model proposed by (Kim and Kim, 2008) and using the logic-based Benders decomposition approach. First, the given model is reformulated to apply the logic-based Benders decomposition scheme. It discusses how to form the main problem and many subproblems. One interesting fact is that this decomposition forms many subproblems of the same type, which makes us use a multi-processing technique. Since the subproblems contain nonlinear terms, it suggests an algorithm using constraint programming technique, whose application generates logic-based Benders cuts that are enforced in the main problem in successive iterations. Finally it ends with a computational results and a discussion of limitations in this study.

product design, logic-based Benders decomposition, constraint programming, conjoint analysis