| With the rapid development of urban rail transit,the metro has developed particularly rapidly in urban rail transit and has become the main mode of transportation to cope with the growing demand for urban travel.At the same time,the passenger flow choosing the metro for travel has sharply increased,and the routes for metro circle line travel are more flexible and complex.Therefore,it is of great significance for the path selection of metro circle line to conduct statistical analysis based on metro circle data and study the passenger flow characteristics of metro circle line.It can not only provide reference value for passenger travel path selection in metro circle line,but also provide scientific theoretical basis for metro operators to dynamically allocate passenger flow on metro circle line.This study uses AFC(Automatic Fare Collection)and AVL(Automatic Vehicle Location)data in the metro circle as research data,and is based on certain data processing and statistical analysis,statistically infers the travel paths of individual passengers,in order to study the choice behavior of metro circle passengers’ travel paths;infer the probability of each passenger riding on each available train using a mixed probability model,and estimate the model parameters using the Expectation Maximization(EM)algorithm;construct a new passenger-to-train allocation model to infer the probability of each passenger riding on each available train.The main results are as follows:(1)Based on the metro circle line without transfer,the metro circle line with one transfer,and the metro circle line with each transfer,an AFC data model and a mixed probability model are constructed,then the maximum likelihood estimation of model parameters is solved using the EM algorithm.(2)Based on different models constructed in three different scenarios,the maximum likelihood estimation values of input parameters are used to construct simulation experiments using certain basic assumptions to test the mean square error(MSE)between the parameter estimation values and the simulated true values,in order to verify the effectiveness and scientificity of the model.(3)The estimated values of the model parameters are applied to the passenger-to-train allocation model established in this article.The input values are only the passengers’ entry time,exit time,and the probability of passengers taking available trains on each route.The route selection results for three different metro circle scenarios are obtained,and the corresponding Beijing metro circle line data is used for example analysis based on the three different scenarios,analyze the distribution of passenger travel time on the corresponding metro circle section and the results of metro circle route selection. |
