| The urban road network is easily disturbed by typical factors such as congestion and random factors such as extreme weather disasters during the operation process,which reduces the road network’s efficiency and capacity.Introducing the concept of service resilience into transportation and urban construction,identifying the characteristics of service resilience of urban road traffic system,measuring static and dynamic service resilience of urban road network,accurately evaluating and predicting road traffic state,formulating refined and flexible congestion prevention and control strategies,and improving the post-disaster recovery ability of the road network are the inevitable requirements for the construction of networked,collaborative and intelligent urban transportation.This paper combines theories and methods such as complex network modeling,data mining,percolation theory,traffic flow theory,deep learning algorithm,and multi-objective optimization decision algorithm.According to the idea of " Assessment of service resilience-Classification of road congestion risk-Prevention and repair strategies ",the traffic congestion control strategy of urban road networks oriented to service resilience is studied.The details are as follows:(1)Static and dynamic service resilience assessment of urban road traffic system.Firstly,the service resilience evaluation framework of the urban road transportation system was clarified,and a hierarchical Bayesian network model was constructed to quantify the static service resilience of the urban road transportation system from the perspective of the system.Secondly,the percolation theory was used to excavate the dynamic relationship between the structure and traffic state of the urban road network,calculate the network’s phase transition threshold,and determine the minimum required performance.Finally,six indicators were proposed to evaluate the dynamic service resilience of the urban road network in the recovery process.(2)Traffic congestion classification of urban roads oriented to service resilience.Firstly,based on the minimum required performance of the road network,a service resilience-oriented evaluation index of urban road traffic congestion was proposed,and the critical threshold of road network congestion diffusion was identified.Secondly,the percolation theory is used to identify the global connectivity level of the road network key road sections.Thirdly,based on the Graph Convolutional Network(GCN)and Gated Recurrent Unit(GRU),a fusion deep learning model is constructed to predict the future evolution trend of traffic congestion.Finally,considering the importance level and congestion index of urban roads comprehensively,the risk matrix is drawn and the congestion level of roads is determined.(3)Research on urban traffic congestion management strategy based on road travel reservation strategy.Firstly,the road capacity distribution function was estimated based on the product limit method.Secondly,the roads’ optimal reservation capacity was determined using the sustained flow index.Then,a Bi-level programming model is constructed.The upper layer model analyzes the road network travel demand based on the optimal reservation capacity constraint,and the lower layer allocates the road network traffic flow.Finally,the Kriging model and the general projection algorithm were used to solve the Bi-level programming model,analyze the traffic state of the road network under the constraint of the optimal reservation capacity,and verify the rationality of the optimal reservation capacity and the effectiveness of the strategy.(4)Emergency repair strategy for post-disaster road networks based on service resilience optimizationFirstly,the evaluation index of comprehensive service resilience of urban road networks was proposed by considering the resilience of road network recovery performance and recovery time.Secondly,the Bi-level programming model for urban road emergency repair was constructed.The upper layer model was a mixed integer programming model considering the resilience of recovery performance and resilience of recovery time.The lower layer model constructed a dynamic traffic flow allocation model based on daily travel demand by considering the behavioral changes of travelers at different periods.Finally,the improved Grey Wolf optimization algorithm and Frank-Wolfe algorithm were used to solve the Bi-level programming model.This study comprehensively considers the traffic state of the urban road network under normal and abnormal conditions,studies the service resilience-oriented urban road traffic congestion control strategy,balances the contradiction between traffic supply and demand,improves the balance of traffic flow in the road network,ensures the service capacity of the road network,improve the resilience of the urban road network after the disaster,and enhance the ability of the road network to withstand various risks.It will support the construction of networked,coordinated,and intelligent urban transportation. |
PDF Full Text Request