An optimization-simulation approach for locating warehouses in logistics networks
This thesis introduces a hybrid optimization-simulation model to locate international and regional warehouses for importer companies. The model assumes multiple products, one or two ports of entry, several geographic regions for the continental US, two international warehouses (IWH) and six regional warehouses (RWH). Demand and transit times are assumed probabilistic and several inventory control mechanisms are assumed. Certain cost based constraints for warehouse locations are included in the
... re included in the model. We conducted many experiments and our observation has been that the warehousing location assignments follow a branching pattern starting from the port of entry and stretching towards downstream echelons. The segments of these branches, meaning distance between Port-IWH or IWH-RWH pairs, get longer or shorter depending on the transportation cost inputs of the model. Basically, in order to minimize the overall transportation costs, the optimization model searches for a solution to balance the cost ratio between these segments. According to our model, warehousing cost and demand distribution appear to be important factors for selecting warehouse locations in the logistics network. If there are multiple potential neighboring regions, the optimization model normally chooses the low-cost region. However, when these regions are similar to each other according to the warehousing cost, the model seeks for a location which is closer to the high-demand regions.