基于复杂网络理论的水网节点重要性评价研究  

Evaluate the node importance for water network based on complex network theory

作　　者：刘忠华[1] 于华[1] 杨方廷[2] 

机构地区：[1]中国科学院大学工程管理与信息技术学院,北京100049 [2]经济领域系统仿真技术应用国家工程研究中心,北京100854

出　　处：《中国科学：技术科学》2014年第12期1280-1294,共15页Scientia Sinica(Technologica)

基　　金：国家重点基础研究发展计划("973"计划)(编号:G2011CB706901);山西省水利厅基金(编号:LZC08-003)资助项目

摘　　要：复杂水网节点的重要性排序在工程应用中有重要实践意义,对于水网调度、安全管理有重要借鉴意义.为此,本文在复杂网络理论基础上,以山西大水网节点重要性排序问题为背景,对水网节点重要性评价进行研究.首先从整体角度出发,不考虑水网方向和干支斗毛级别差异,基于水网全局网络拓扑结构,从局部属性、全局属性、传播属性和网络位置四个方面定义了节点重要性评价指标.其次,分别根据这四个指标,对水网节点重要性进行排序分析,结合分析结果对网络位置指标进行了改进.然后,由于单指标评价结果不够全面且适用范围具有很大的局限性,给出了基于TOPSIS的多属性决策方法来综合评价节点的重要性.最后,考虑水网方向和级别差异,提出对水网进行干支斗毛级别划分,基于Page Rank算法对同一级别的有向赋权水网节点重要性评价进行研究.There is a significant application to evaluate the node importance of water network in water resources schedule and safe operation management. With the background of sorting the nodes of Shanxi Water Network, this paper studies to evaluate the node importance for water network based on complex network theory. Firstly, four indexes are defined from the network structure and transmission dynamics perspective, such as Degree Centrality, Closeness Centrality, Betweenness Centrality and K-Shell. Secondly, this paper improves one index and proposes a multi-attribute decision method to evaluate the node importance more comprehensive. Finally, with consideration of the direction and level in water network, the water network should be divided into several sub-networks with the same level, and the node importance in the same level network is studied based on PageRank algorithm.

关 键 词：节点重要性 复杂网络 多属性决策 PAGERANK算法 山西大水网 

分 类 号：TV213.4[水利工程—水文学及水资源]