Rail Transit Short Term Passenger Flow Forecasting And Transfer Coordination Study

In recent years,our country’s urban rail transit developed rapidly on the basis of the original,and the urban rail transit gradually formed a scale.Rail transit has become an essential way that people travel.The requirements for urban public transport service are also getting higher and higher.Therefore,it is necessary to study the method of real-time and accurate prediction of rail transit passenger flow and the model of transfer coordination evaluation.In this paper,the short-term passenger flow of Beijing metro line 14 and the transfer coordination between the stations and the ground buses are studied systematically,including the analysis of the characteristics of passenger flow,the real-time passenger flow forecasting of artificial neural network model,the evaluation,analysis and suggestion of transfer coordination between rail transit and ground bus.The main research work includes:(1)Distribution characteristics of short term passenger flow in time and space are introduced in detail,and Beijing metro line 14 stations are classified into different passenger flow characteristic types based on the actual data of passenger flow.Accurately grasp the characteristics of the 14 lines of the station passenger flow,to lay the foundation for accurate prediction.In order to achieve the purpose of real-time prediction,the prediction time is divided into 5min.So that,the model need good anti noise ability to adapt to the high nonlinearity and oscillation of the passenger flow distribution.Therefore,the wavelet neural network is used as the prediction model.Design the network topology model and adjust the parameters of the model according to the sample data analysis results.(2)Aiming at the problem that the wavelet neural network may be trapped in local minimum and can not get the global optimal value,genetic algorithm and immune algorithm are used to optimize the model.Based on the prediction results of different passenger flow characteristics of Beijing metro line 14,use the model evaluation index to compare the prediction results of wavelet neural network.It is proved that the optimized network model has the characteristics of global optimization,higher accuracy and equalization coefficient.The model has the advantage of short term passenger flow forecasting.(3)Evaluate the transfer situation of rail transit and ground bus.Select the appropriate evaluation index and establish the evaluation system with nine indexes.Matter element analysis method based on extension theory is used to design the evaluation model,and the genetic algorithm is used to calculate the model.Use the model to evaluate Jiangtai Station and Zaoying Station of Beijing metro line 14,analyze the results and give reasonable suggestions.