A stochastic programming approach to the location of distribution centers for multinational enterprises under demand uncertainty

Abstract

Multinational enterprises (MNEs) often collaborate with local agents to establish initial distribution channels due to their need for market-specific knowledge and experience. As the market matures and upstream suppliers and production plans are solidified, MNEs may transition to developing their distribution systems and supply chain networks. Integrating the transportation network among upstream material suppliers, production facilities, and distribution centers (DCs) becomes crucial at this stage. Since transportation costs constitute a significant portion of enterprise expenses, optimizing upstream transportation is essential for MNEs following this market entry strategy. This study aims to optimize the location decisions of DCs while assuming that suppliers, plants, and retailers have fixed locations. A critical focus is the integration of upstream transportation operations, specifically between suppliers and plants and between plants and DCs, to minimize inefficient empty backhauls. Additionally, demand uncertainty is factored into this long-term strategic design problem. A stochastic programming (SP) model is developed, and a solution procedure based on the Genetic Algorithm (GA) is designed to handle practical-scale problems. Numerical experiments demonstrate that the GA method achieves a solution quality with less than a 1 % gap compared to the optimal solution while also significantly reducing computation time.