Simulation-based Research On Dynamic Passenger Flow Distribution In Urban Rail Transit Networks

With the progress of urban rail transit construction,the system structure of urban rail transit network(URTN)becomes more and more complex,and the traffic increases sharply.Accurate prediction of passenger flow distribution and the use of various transport equipment are the guarantee of efficient resource allocation and orderly network operation.Sections and time periods are subjects in traditional study of passenger flow assignment.Traditional study cannot meet the needs of the operator for more precise and dynamic information about passenger flow.Adopting the multi-agent modeling,passengers,trains and stations as three main elements of the system are analyzed.And the construction method of URTN simulation model is presented in this paper.Real-time information of passengers,trains and stations in the networks is recorded and displayed,which is conducive to study the rule of passenger flow distribution in the time dimension.The detailed research contents include:(1)Analysis of passenger travel behavior.This paper cleaned AFC data by Python.The characteristics of travelers in URTN is studied based on AFC data.The topology of URTN and solution of effective path is discussed.A traveling model considering boarding process as basic unit is presented.(2)Analysis of station service process and train service planning.Service time fitting method with Weibull function is proposed to describe the time distribution of travelers’ behavior in station.Simulation models based on typical station is constructed to sample and test the accuracy of the fitting method.The basic characteristics of URTN traffic are analyzed to determine the organization of train.(3)Construction of the URTN simulation framework based on AFC data.The characteristics of the subjects in the URTN is analyzed.An agent-based simulation module is constructed.At last a Multi-Agent simulation system for URTN is proposed.(4)Simulation modeling with the case of Chengdu metro system.This paper designs the simulation system and the UI interface programed by Anylogic.5% AFC data of a certain day is taken as the input data to verify the simulation system.The result shows that the travel distribution error is only 9.1% compared with the actual AFC.The accuracy and rationality of the model are proved by analyzing some of the passenger flow rules output by the simulation system.