| The uncertainty of traffic demand poses significant challenges to the identification of critical links and the transportation network design,resulting in poor identification and optimization results.Macroscopic Fundamental Diagram(MFD)can avoid this problem,it is an intrinsic property of network and insensitive to traffic demand.MFD is often used as a tool to evaluate the operational status of network and network performance.Therefore,in order to avoid the interference of the uncertainty of traffic demand,based on MFD theory,this paper proposes a critical links identification method,and constructs the discrete transportation network design model.The proposed method and model are verified by designing simulation experiments.The main research contents are as follows:(1)A method of regional network MFD modeling based on SUMO simulation is proposed.Taking the weighted traffic and weighted density of network as the parameters of MFD,SUMO simulation was used to establish the local network in Yanta District,Xi ’an City.At the same time,relevant simulation environment and parameters were configured to obtain the simulation network data and draw the scattered points of MFD,which verified the existence of MFD.The MFD data were fitted and analyzed,and the critical points of MFD were extracted.The Gaussian mixture clustering algorithm(GMM)is used to classify the traffic state and identify the threshold of traffic state classification,which lays a foundation for the subsequent research.(2)Proposed the identification method of critical links of the network based on MFD.The influence of links failure on MFD was used as the evaluation basis for critical links.Based on the influence mechanism of links failure on MFD,the evaluation indexes of MFD critical point offset and the difference in MFD passing flow were constructed.At the same time,considering the influence of links failure on network users’ travel behavior,a network flow update method based on UE user equilibrium is proposed.According to MFD critical point offset and the difference in MFD passing flow,a comprehensive index of links criticality was constructed,and the links criticality was ranked based on the index,so as to divide the critical links.(3)A discrete network design model based on MFD and critical links is constructed.The vertical coordinate corresponding to the critical point of MFD(i.e.the maximum weighted flow of MFD)was taken as the evaluation index of the road network performance.At the same time,the identified critical links were taken as the alternative links of the network design.A Bilevel programming mode of discrete network design based on lane extension was constructed.The objective of the Upper-Layer model is to maximize the maximum weighted flow of MFD and to find the optimal network performance scheme among the feasible network design schemes.The objective of the Lower-Level model is to update the network flow through the UE user equilibrium model,so as to update the MFD.The model is calculated and implemented through simulation traversal,that is,through the simulation traversal of each feasible network design scheme,so as to obtain the network performance index corresponding to each scheme,and then select the best scheme.(4)The proposed method and model are verified by SUMO simulation.Several types of critical links in the network are identified through SUMO simulation traversal.At the same time,the discrete transportation network design based on lane extension is carried out according to the obtained critical links.The optimal scheme is selected from the feasible schemes through simulation traversal,and the scheme is evaluated from the perspective of the stability of the network status and the engineering economy.The results show that the proposed method based on MFD can accurately identify critical links in the network,and the discrete transportation network design based on lane extension for the critical links can effectively improve the traffic performance of the network. |
