基于空间聚类的我国应急物资储备区域划分实证研究

On China's emergency resources storage region division based on the spatial clustering

通过对国内外应急管理区域化发展模式总结,提出我国区域化应急物资储备体系建设的必要性。利用空间聚类方法,以储备成本和救援效率为目标建立数学模型,并给出动态聚类算法,以解决应急物资储备区域划分问题;最后对我国应对自然灾害的应急物资储备区域划分进行了实证研究。

The present paper takes it as its research goal to make an ideal regional division for the emergency resource storage and management in our country. As is known, in recent years, it has become one of the key influential factors to store and manage emergency resources properly and effectively to meet the demands for storage and delivery of emergency goods and resources, which may influence and decide the success of the disaster relief and the rational distribution and delivery of emergency goods. Analyzing the emergency resources management and delivering experience from abroad let us know that the regional management of emergency goods has become one of the chief management patterns in the USA, Britain, Canada and other developed countries. Meanwhile the emergency resources reservation system should be taken in the regional mode so as to improve the efficient utilization of such resources and reduce the cost for their storage and distribution. At the same time, it should be more convenient and appropriate for each region to make the regional emergency response plan and designate its emergency resources storage location and allocation according to its own regional disaster-incidence characteristic features, which may become the primary and key problem for the urgent issue or issues for the place in a critical moment. In this paper, we have first of all analysed the feasibility of the spatial clustering method to solve the＇division problem. Then, we have set up an emergecy resources region division model by using the spatial clustering method to confirm the new clustering objective function to keep the balance the respective regions from the point of view of such resources storing and transportation cost and the scope of the different regions （including the rescuing missions）. And, next, we have put forward an improved dynamic clustering algorithm based on the kmeans clustering algorithm to solve the emergency resources division problem in the best way in terms of data extraction, the region division of natural disasters in our country in accordance with the proper regional division strategies. The rationality of the given region division results and the optimum region-division number has been analyzed so as to promote the solution of the problem in an ideal manner in accordance with the actual situation of China today.