考虑时序动态和恢复决策的高速铁路网络韧性研究OA

To effectively evaluate the performance of high-speed railway transport networks under sudden disruptions and develop scientifically grounded recovery strategies,this paper proposed a resilience assessment framework for high-speed railway networks that incorporates temporal dynamics and recovery decision-making.First,the temporal variations in network structure and operational states induced by heterogeneous train flows were analyzed.A high-speed railway temporal transport network model was constructed based on temporal graph theory,along with a formula for calculating the average temporal path length.A network performance response function was established by integrating passenger travel time and travel intensity.By using the network performance retention rate as the resilience metric,this study relaxed the conventional assumption of"sequential repair by a single repair team"in network recovery and extended it to"joint repair by multiple repair teams".A Markov decision process was employed to model recovery decision-making under multiple disturbances,and a Q-learning-based reinforcement learning algorithm was designed to derive the optimal recovery strategy that maximizes resilience.Based on the topological structure and operational datasets of China's high-speed railway network,the network resilience under different failure and interruption scenarios at stations and sections was investigated.The results indicate that the resilience of the high-speed railway network exhibits pronounced spatiotemporal heterogeneity.Under functional failure scenarios,stations with greater impact are discretely distributed at the intersections of main lines,with the network showing the lowest resilience level at 15:00.Under facility disruption scenarios,stations and sections with greater impact are concentrated along a"two vertical and one horizontal"corridor,with the most severe network performance loss occurring at 6:00.The importance rankings of stations and sections vary over time and are not constant.In disruption scenarios,the importance of stations or sections demonstrates a stronger correlation with temporal betweenness.Furthermore,multi-section disruptions occurring in hub regions lead to significantly greater network performance degradation than those in random regions.Compared to four other recovery strategies,the optimal recovery strategy based on Markov decision-making under flexible constraints enables flexible repair scheduling,improves network resilience by over 2%,and allows the recovery curve to reach a higher resilience level more rapidly.The research findings can provide decision-making support for the operation and maintenance of high-speed railways.