城市轨道交通站点失效修复策略

为提高城市轨道交通（urban rail transit，URT）系统应对突发事件的能力，研究了URT站点在自然灾害及人为破坏情况下的修复策略. 分别运用随机攻击及蓄意攻击模拟URT站点遭受自然灾害及人为破坏的情况，结合仿真思想提出失效URT网络的仿真修复策略，运用平均修复策略、偏好修复策略及仿真修复策略实现受损URT网络的修复，结合全网可达性及未受影响乘客占比，提出全网可达性损失和全网未受影响乘客占比损失两类系统弹性损失指标用于衡量修复策略的有效性. 结果表明:蓄意攻击情况下全网可达性及未受影响乘客占比分别下降了2.54%及64.82%，高于随机攻击的2.34%及55.28%，蓄意攻击对URT网络的损害更大；修复因随机攻击而失效URT网络时，仿真修复策略的两类弹性损失指标分别为0.009及0.182，优于平均修复策略的0.012及0.305和偏好修复策略的0.010及0.197，修复因蓄意攻击而失效URT网络时，仿真修复策略的两类弹性损失指标分别为0.009及0.258，优于平均修复策略的0.012及0.312和偏好修复策略的0.014及0.354，表明仿真修复策略更适用于完成受损URT网络的修复工作；修复全程中仿真修复策略曲线的平均斜率为0.146，大于平均修复策略的0.092及偏好修复策略的0.117，表明相比于常规修复策略，仿真修复策略能达到更高的修复效率；当维修资源受限时，采用仿真修复策略能够达到最优的修复效果. 

Repair Strategies for Failure of Urban Rail Transit Stations

In order to improve the ability of urban rail transit (URT) systems to respond to emergencies, the effective repair strategies for URT stations under natural disasters and human attack are discussed. Firstly, the station failure caused by natural disasters and human attack are simulated by random attack and intentional attack, respectively. Secondly, the simulation repair strategy for damaged URT network is proposed. Finally, the average repair strategy, preference repair strategy and simulation repair strategy are used to repair the damaged URT network, and the effectiveness of repair strategies is measured with two categories of resiliency loss indicators: global accessibility resiliency loss and global proportion of unaffected passengers resiliency loss, which are proposed on the basis of the global accessibility and proportion of unaffected passengers. The results show that the global accessibility and proportion of unaffected passengers decrease by 2.54% and 64.82% in case of intentional attack, which are higher than 2.34% and 55.28% of random attack. Intentional attack is more harmful to the URT network than random attack. When the damaged URT network caused by random attack is repaired, the resiliency loss indicators of the simulation repair strategy are 0.009 and 0.182, which are better than 0.012 and 0.305 for average repair strategy and 0.010 and 0.197 for preference repair strategy; when the damaged URT network caused by intentional attack is repaired, the resiliency loss indicators of the simulation repair strategy are 0.009 and 0.258, which are better than 0.012 and 0.312 for average repair strategy and 0.014 and 0.354 for preference repair strategy. It indicates that the simulation repair strategy is more suitable for repairing the damaged URT network. The average slope of the simulation repair strategy curve is 0.146, which are higher than 0.092 for average repair strategy and 0.117 for preference repair strategy. Compared with conventional repair strategies, the simulation repair strategy has higher repair efficiency. Using simulation repair strategy can achieve optimal result when the repair resources are limited.