考虑引力关联的机场网络节点重要度分析

Node importance analysis of airport network considering gravitational correlation

机场网络中关键节点的失效会对网络运行造成一定影响。准确识别关键节点可提升整个网络运行的安全性、正常性和高效性,同时为基于空中交通网络的大面积航班延误治理提供一定的决策依据。基于万有引力定律,以节点间的引力构建具有引力关联的机场网络,对传统的邻域结构洞指标进行改进并结合复杂网络选取适当的评估指标;结合指标之间的差异和冲突,采用基尼系数、肯德尔系数来计算权重,以此建立评估指标体系。同时,引入弹性系数并结合改进的前景理论,得到前景值,采用以相对熵代替欧氏距离的TOPSIS方法计算节点的综合度,将前景值和综合度结合得到节点的重要度。另将节点延误状态作为一种失效模式,建立考虑延误波及的级联失效抗毁性模型,量化网络抗毁性。最后,以中国东北、华北、华中、华东部分机场网络进行实例对比和仿真分析。分别以所提方法和其他3种方法评估该网络节点的重要度,并各选取排名前五的节点计算网络抗毁性。结果表明,所评估出的北京、郑州、青岛、大连、太原节点对于整体网络的重要性最高,一旦发生延误对网络的抗毁性影响最大。

The failure of any key node in the airport network will cause a certain influence on network operation.On the contrary,identifying these key nodes can accurately improve the operation safety,normality,and efficiency of the entire network,and also provide a certain decision-making basis for the management of large-scale flight delays based on air traffic networks.Based on the law of gravitation,an airport network related to gravity was established in this paper by taking gravity between nodes.The traditional neighborhood structure hole index was improved and appropriate evaluation indexes were selected based on the complex network;Gini coefficient and Kendall coefficient were also used to calculate the weight by combining the differences and conflicts between indicators and an evaluation index system was also established based on the above.Besides,the prospect value was obtained by introducing the elastic coefficient and combining the improved prospect theory and the comprehensive degree of nodes was calculated by the TOPSIS method where K-L divergence replaces Euclidean distance.Finally,the importance of node was obtained by combining prospect value and comprehensive degree.In addition,a cascading failure invulnerability model that considers delay propagation was established by taking node delay status as a failure mode.The delay load of node transmits via airport network.During the process above,the failed points,the affected points,and the affected edges were generated and used to quantify the invulnerability of the network.Finally,example comparison and simulation analysis were made on some airport networks in Northeast China,North China,Central China,and East China.The importance of network nodes is evaluated using the method in this paper together with the other three methods and the top five nodes in each method were selected to calculate network invulnerability,respectively.As indicated by the results,Beijing,Zhengzhou,Qingdao,Dalian,and Taiyuan nodes evaluated in this paper are the most important ones to the overall network,and delays in these airports exert the greatest influence on network invulnerability.