A novel bi-level programming model for facility location in supply chain network design

Facility location models can be considered as the foundation of supply chain network design models that deal with the location of facilities in some given geographical area. Basic facility location problems used for supply chain management often consider a single product and a single production / distribution echelon with uncapacitated or capacitated facilities. Many of them are formulated based on the assumption that the number of candidate facility sites in the supply chain is given. Various models developed for maximal coverage have been utilized to obtain the candidate sites in the given customer demand area. This paper focuses on the problem of selecting locations for distribution centers or retailers of a supply chain to maximize the total customer demand area covered by these facilities while at the same time keeping the distance between facility sites as far apart as possible. A bi-level programming model is constructed with the consideration of maximizing the distance between each pair of selected facilities to solve the proposed problem. Genetic Algorithm is applied to solve the model. Numerical results of a case study indicate that the proposed model is more realistic and can be readily embedded into the supply chain network design.