Research On Prediction And System Of Distribution Of Waiting Passengers On Subway Platforms Based On Ant Colony Algorithm

With the constant expansion of the scope of the rail transit network and the increasingly improved network operations in major cities,passenger flow has increased significantly,and the congestion problem of passenger flow on station platforms has become increasingly serious during the morning and evening rush hours.In the case of controlling the passenger flow into the station,there is still passenger congestion in some parts of the platform.The essential reason is the uneven distribution of waiting passengers on the platform.This disequilibrium is mainly reflected in the larger number of passengers in some waiting areas while the number of passengers in other waiting areas is relatively small.Therefore,studying how to predict the disequilibrium in the distribution of waiting passengers on a platform is of great significance for platform operation managers to formulate passenger flow guidance measures.On the basis of studying the characteristics of subway platform passengers and the passenger distribution in the platform waiting area,this paper constructs a platform waiting passenger distribution prediction model,and studies the prediction of platform waiting distribution during the peak passenger flow period.The main research work is as follows:(1)Research on the characteristics of subway platform passenger flow and the passenger distribution in the platform waiting area.Firstly,the characteristics of passenger flow in subway platform are analyzed from two aspects: the travel time and space characteristics of passengers and the characteristics of passengers’ waiting behavior.Secondly,the content and method of data survey are determined,and representative stations are selected according to the classification of stations for field survey to obtain data.Finally,according to the results of the survey,the distribution of waiting passengers on various platforms is analyzed.(2)Analysis on influencing factors of platform waiting passengers’ choice behavior.From the perspective of the selection behavior of waiting passengers,the influencing factors that lead to the uneven distribution of waiting passengers on the platform are analyzed,including relative walking distance,waiting position and queuing situation,the position and number of escalators in the platform building,passenger arrival volume,platform passenger flow guidance measures,and passengers’ own factors.(3)Construction of a prediction model for the distribution of waiting passengers on the platform based on the ant colony algorithm.According to the analysis results of factors influencing the selection of waiting positions by waiting passengers on the platform,the calculation of state transition probability in ant colony algorithm is improved,and the probability function of passenger waiting position selection with relative travel distance and waiting position queuing status as independent variables is established.On this basis,a prediction model for the distribution of waiting passengers on the platform is constructed with the goal of minimizing the relative error of the number of waiting passengers on the platform and introducing the constraint condition of the maximum number of waiting passengers at the waiting position,and the model algorithm is designed to solve the number of passengers waiting at each station with a certain number of passengers arriving.(4)The design of the urban rail transit platform waiting distribution prediction system based on the platform waiting passenger distribution prediction model.Taking Xizhimen Station Line 2 station platform in the direction of Chegongzhuang as an example,the feasibility and accuracy of the model are analyzed and verified.Based on the My SQL database,system programming is carried out through Visual Studio 2019 to realize the system’s platform waiting passenger distribution prediction function and waiting distribution information display function.There are 59 pictures,17 tables and 50 references.