Enhancing facility location modeling and optimization

点击翻译按钮获取中文摘要

One of the most important ways an industrial firm, retail outlet, or government agency can enhance its chances of success is to identify good facility locations. One approach is to use location-allocation models that optimally locate service or manufacturing facilities and allocate demand or supply to them: the intent is to find the best set of locations and service areas for these facilities. Location/logistics models are defined by their objective functions and what are thought to be realistic constraints. However, the objective functions and the constraints in the classic models tend to be quite simple. Consider, for instance, the most popular location-allocation model, the p-median, where the objective is plainly to minimize the distance (or travel cost) between demand and facility points. In many real-world situations, assumptions as uncapacitated facilities, demand points that are allocated to the closest facility (traditionally a constraint in the optimization models), and point-to-point service are unrealistic. Relaxing or removing some of these assumptions is critical to enhancing facility location modeling and optimization in an attempt to generate solutions that better incorporate real-world conditions.