地理科学 ›› 2013, Vol. 33 ›› Issue (2): 143-149.doi: 10.13249/j.cnki.sgs.2013.02.143

• • 上一篇    下一篇

基于复杂网络聚类的最优选址模型

戴技才1(), 宗会明2()   

  1. 1.重庆师范大学地理与旅游学院 重庆 400047
    2.西南大学地理科学学院,重庆400715
  • 收稿日期:2012-01-09 修回日期:2012-04-09 出版日期:2013-02-20 发布日期:2013-02-20
  • 作者简介:

    作者简介:戴技才(1974-),男,湖南益阳人,博士,讲师,主要从事地理信息系统模型,复杂系统研究。E-mail:daijicai@cqnu.edu.cn

  • 基金资助:
    教育部人文社会科学研究青年基金(批准号: 11YJCZH023);国家自然科学基金项目(41271411,41001102)资助

Model of Locating Optimal Sites Based on Complex Network Clustering

Ji-cai DAI1(), Hui-ming ZONG2()   

  1. 1.College of Geographical Sciences, Chongqing Normal University, Chongqing 400047, China
    2.School of Geographical Sciences, Southwest University, Chongqing 400715, China
  • Received:2012-01-09 Revised:2012-04-09 Online:2013-02-20 Published:2013-02-20

摘要:

最优选址在社会经济活动中非常重要。传统的网络聚类分析以空间两点之间的直线度量距离,而不是以空间最短网络路径作为聚类条件,无法找到复杂网络的最优选址中心。基于最短路径的复杂网络聚类模型,探索复杂道路网络中的最优选址。模型通过迭代法获取近似最优解,二分邻域分割法逼近最优解分布区,应用邻域下降法达到最优选址点。实验结果表明:本模型与穷举-Dijkstra算法相比,计算精度相当,计算速度提高了约23倍以上。模型以复杂网络聚类为基础推导,为复杂网络选址、聚类提供了一种新的理论与方法。

关键词: 选址, 网络聚类, 复杂网络, 最短路径, Dijkstra算法

Abstract:

The best location for site selection is very important for social and economical systems. Traditional clustering analysis cannot reach the optimal location allocation center based on the Euclidian, Manhattan, Chebyshev, Fu or Tanimoto distance,which does not depend on the shortest path distance as the clustering condition. Complex network clustering model based on the shortest path is proposed to explore site selection in this paper. The approximate optimal results are obtained using the iterative technology. The searching range of optimal site allocation is approached by half dividing the neighborhood. And the best location is captured applying the descend strategy. Experiments indicate that the proposed model can improve the calculating speed by at least 23 times with the same precision to approach the optimal resolution compared to the greedy Dijkstra algorithm. This model is inferred from the complex network clustering. It provides new theory and method to cluster complex network and locate sitting facilities.

Key words: location allocation, network clustering, complex network, shortest path, Dijkstra algorithm

中图分类号: 

  • K903