容量可靠性视角下的货物运输网络关键枢纽识别方法

Critical hub identification method for freight transportation networks from a capacity reliability perspective

货物运输网络部分关键枢纽被随机或蓄意攻击后,常会出现级联失效现象,导致网络容量可靠性下降,影响网络正常运营.因此,本文通过改进的有向含权度中心性、有向含权介数中心性和有向含权LeaderRank算法识别货物运输网络中的关键枢纽.同时,进一步构建货物运输网络级联失效模型,从容量可靠性视角对不同算法识别的关键枢纽进行随机和蓄意攻击模拟仿真,确定有效的关键枢纽识别方法.仿真结果表明:有向含权度中心性可有效识别对网络容量可靠性影响大的关键枢纽;容量系数和过载系数的提升能有效保证网络容量可靠性且存在突跃现象;失效或过载节点进行负载分流时,相比于考虑业务关系紧密程度,考虑下游节点的冗余容量能有效减少级联失效,提升网络容量可靠性.研究结果对货物运输网络预防级联失效,保证网络容量可靠性具有重要借鉴意义.

Cascading failures often occur when some critical hubs in cargo transportation networks are attacked randomly or deliberately,leading to a reduction in network capacity reliability and affecting network operations.In this study,we use the improved weighted and directed degree centrality,weighted and directed betweenness centrality and weighted and directed LeaderRank algorithms to identify critical hubs in freight transportation networks.In addition,we further construct a cascading failure model for freight transportation networks,and simulate random and deliberate attacks on the critical hubs identified by different algorithms from the perspective of capacity reliability,so that an effective critical hub identification method can be determined.Our simulation results show that the weighted and directed degree centrality can effectively identify critical hubs that exert a significant impact on network capacity reliability.Furthermore,the increase of capacity factor and overload factor can effectively ensure the network capacity reliability and a sudden jump phenomenon occurs.Also,simulation results confirm that considering the redundant capacity of downstream nodes upon load shunting of failed or overloaded nodes effectively(1)reduces cascading failures and(2)improves network capacity reliability,compared to realistically considering the closeness of business relationships.These research results secure important reference significance for preventing the cascade failure of freight transportation networks and ensuring the network capacity reliability.