对空间数据挖掘聚类技术及其在公共设施选址方面的传统应用方法进行了综述,分析了传统应用方法中有待解决的关键问题,对空间距离代价的表示问题和传统方法的算法时间复杂度进行了初步探讨,运用模拟退火算法和图论对传统方法进行了改进,实现了算法时间复杂度的降低和聚类结果的优化。 This paper summarizes the spatial data mining clustering technology and its traditional application in the public facilities location, analyzes the key problems to be solved in the traditional application methods, analyzes the spatial distance cost representation and the traditional algorithm time complexity A preliminary study was carried out. The traditional methods were improved by simulated annealing algorithm and graph theory, which reduced the time complexity of the algorithm and optimized the clustering result.