重力p-median模型在设施选址中的应用及检验

Application and validation of gravity p-median model in facility location research

区位分配模型是设施选址研究中的重要方法,其中p-median模型是应用最广泛的一种.但传统p-median模型中每一需求点仅选择最近的那一个设施的假设具有一定局限性.重力p-median模型对此进行了重要改进,假设需求者根据距离和设施吸引力确定选择各个设施的概率.但目前学界关于重力p-media:n模型在实际应用中的有效性仍未达成一致.本文将重力p-median模型应用到北京市延庆县医疗设施布局的实际案例中,并与p-median模型进行比较,以检验重力p-median模型的有效性和适用范围.结果表明,重力p-median模型作出的改进具有实际意义,当设施候选点吸引力存在差异时,重力p-median模型会得到与传统p-median模型显著不同的优化结果;两个模型适用于不同的设施选择行为,在实际应用中需把握清楚所研究设施的特点才能做出合理的选择.

Location allocation modelling is an important method in facility location research. Among the various methods in location allocation modelling, the p-median model is the most widely used. However, one of the limitations of p-median model is its assumption that each demand point only accesses to the closest facility. The gravity p-median model makes an important improvement by assuming that the service users access to facilities depending on the distance and attraction of each facility. A consensus, however, has still not been reached among researchers. This study applies the gravity p-median model in the case study of health care facilities in Yanqing county, Beijing, and compares its results with the results of the p-median model, aiming to empirically test the actual validity and applicable cases of the gravity p-median model. The results show that the improvements made by gravity p-median model to p-median model are significant, especially when the attraction of facilities significantly varies. The two models are suitable for different scenarios. Researchers should choose a suitable model depending on the fully understanding of the utilization behavior of services.