Optimal design of the auto parts supply chain for JIT operations: Sequential bifurcation factor screening and multi-response surface methodology

This research proposes a solution framework based on discrete-event simulation, sequential bifurcation (SB) and response surface methodology (RSM) to address a multi-response optimization problem inherent in an auto parts supply chain. The objective is to identify the most efficient operating setting that would maximize the logistics performance after the expansion of the assembly plant's capacity due to market growth. In the proposed framework, we first construct a comprehensive simulation as a platform to model the physical flow of the auto parts operations. We then apply the SB to identify the most important factors that influence system performance. To determine the optimal levels of these key factors, we employ RSM to develop metamodels that best describe the relationship between key decision variables and the multiple system responses. We adapt the Derringer-Suich's desirability function to find the optimal solution of the metamodels. Computational study shows that our method enables the greatest improvement on system performance. The proposed method helps the case firm develop insights into system dynamics and to optimize the operating condition. It realizes the performance objective of the auto parts supply chain without the need for additional fiscal investment.