Optimization Of Urban Rail Transit Security Resources Based On Short-term Passenger Flow Forecast

With the continuous expansion and improvement of urban rail network in major cities,urban rail transit is increasingly used by residents as the preferred mode of public transportation.In order to ensure the travel safety of residents and the normal operation of the station,a security check area has been set up at the rail transit station.In the case of a large number of passengers entering the station,the problem of passengers queuing and staying in the security inspection area often occurs,and there is a greater safety hazard.This is due to the lack of methods to accurately predict short-term passenger flow changes in stations,the lack of awareness of the queuing mechanism of passenger security check,and the lack of scientific basis for dispatching security check service resources,resulting in extensive management of station security check resources.Therefore,this paper starts from the short-term passenger flow prediction at the inbound station,builds a short-term passenger flow prediction model with high accuracy,starts with the process of the security inspection system,analyzes and studies the method of optimizing the allocation of security inspection resources,and provides a reasonable security inspection service resource allocation and scheduling for the station theoretical support.The paper first introduces the definition and classification of the phenomenon of large passenger flow,the basis for judgment and the method of passenger flow organization when dealing with large passenger flow.According to the spatial and temporal distribution of inbound passenger flow,the characteristics of inbound passenger flow are analyzed.This paper introduces the concept of subway security inspection,the process of security inspection and its influencing factors,and expounds the necessity of optimizing the allocation of security inspection resources in life.This may lay the theoretical foundation for subsequent research.Then,through STL time series decomposition,the inbound passenger flow sequence with high randomness is decomposed into three components,and the complexity between components is compared through sample entropy analysis.The secondary decomposition greatly reduces its randomness,fully decomposes the internal passenger flow information of the margin,and obtains the IMF component with strong regularity.LSTM long short-term memory neural network is used to predict the components obtained by the two decompositions,and the root means square error(RMSE),means absolute error(MAE),means absolute percentage error(MAPE)and coefficient of determination(~2)of the prediction results are used to predict four error evaluation indicators to evaluate the effectiveness of the combined model.According to the error index,it can be proved that the STL-EEMD-LSTM combined prediction model has high prediction accuracy and can be used as a tool for short-term inbound passenger flow prediction.Finally,according to the security inspection process,the security inspection system is divided into three queuing processes,and a queuing network of the security inspection system is constructed.Taking the waiting time of passengers and the total cost of security check as the optimization goals,a multi-objective optimization model of security check resource allocation is constructed.Constructing inbound passenger flow characteristics are based on historical inbound passenger flow data.Using K-means clustering algorithm to divide inbound passenger flow during one day’s operation into periods of different inbound passenger flow levels.The time period is solved by NSGA-II multi-objective algorithm.Through the analysis of an example,the station’s current security check resource allocation is compared with the algorithm solution results to verify the effectiveness of the model in real life,and corresponding resource allocation improvement measures are proposed.