Study On Passenger Flow Assignment Method Of Urban Rail Transit Based On Information Feedback

Urban rail transit is an important part of urban comprehensive transportation system.The passenger flow assignment method of urban rail transit system is the basis of its network upgrading and operation organization optimization.With the continuous expansion of passenger demand in urban rail transit network,congestion and passenger retention are becoming more and more serious,which not only affects the operation safety and system service level,but also affects the passenger flow assignment characteristics of the network.Therefore,this paper studies the urban rail transit passenger flow assignment method,the main research contents are as follows.1)Basic data processing and analysis.AFC data,train schedules and network data as input data are the basis of passenger flow assignment.Passenger OD and actual travel time are obtained by combining AFC data as the assignment subjects.The arrival and departure time of each train is determined by combining the train timetable data as the assignment object.Combined with network data,the effective path between ODs is obtained.The path search is carried out on the basis of constructing the topological network.If the transfer time is ignored in the search,the estimation accuracy of the path travel time will be reduced.Therefore,a depth-first search algorithm considering the influence of transfer is proposed.The upper limit of transfer number and travel time are used for branch and bound,so as to quickly determine the k shortest routes between ODs.2)The calculation method of path selection.Determining the passenger selection path is the premise of loading the passenger flow path.When studying passengers’ path selection,few literature considers that passengers update their path decision-making behavior according to historical passenger flow assignment information during travel.In view of this deficiency,a path selection calculation method based on matching AFC data is proposed.The path selection calculation is based on the travel time of the path as the main basis,and the number of candidate path transfer and congestion degree as the secondary basis.Combined with the transfer time and congestion degree information of the first departure passengers,the path attribute is updated,thus affecting the path selection of the second departure passengers.3)Path selection model.The travel time plays a leading role in the result of route choice when using the above method,but the number of transfer and congestion are also important factors affecting the passenger route choice in actual decision-making.Therefore,a route choice model based on semi-compensation theory is proposed to comprehensively consider various factors.The model gives certain randomness to different passenger transfer times and transfer time tolerance,reflecting passenger heterogeneity;path selection probability is determined by path generalized cost function and utility function;according to the transfer time and congestion information of departure passengers,the path attributes are updated,which affects the selection probability of different paths.For the parameters in the model,a calibration method based on Markov chain and genetic algorithm is proposed.In addition,in the path selection model,it is necessary to estimate the walking time and transfer time of inbound and outbound stations.Therefore,a method for estimating the walking time and transfer time of inbound and outbound stations based on AFC data and train departure interval is proposed.4)The construction of assignment model and simulation framework based on information feedback.In the existing assignment model,the train capacity limitation,passenger platform waiting process and transfer process are seldom considered,which makes it impossible to reflect the influence of train capacity limitation on passengers ’ boarding behavior,as well as the change of transfer number and transfer time.Therefore,in the construction of the model,this paper considers the constraint of the maximum passenger capacity of the train,and uses the simulation method to simulate the whole process of passenger travel.By marking each passenger,the transfer time and congestion information of the first departure passenger are obtained,and the path attribute is updated with this information.The simulation framework is built on the basis of the interaction among passengers,trains and stations,starting from the train departure,from the perspectives of determining the departure train and loading passengers to the train.Finally,the assignment accuracy with and without information feedback method is compared to verify the effectiveness of information feedback method,and the operation status of Chengdu rail transit network is analyzed from the aspects of train passenger volume,section flow and transfer number.