Research On Dynamic Scheduling Optimization Of Single-Line Bus Based On Short-Term Passenger Flow Prediction

In recent years,as residents travel in more diversified ways,residents have increasingly high requirements for service levels such as speed,directness and comfort of public transportation.The traditional bus scheduling strategy has problems such as low operational efficiency,long passenger time in the car and insufficient capacity utilization,while dynamic bus scheduling can improve the problems of traditional bus scheduling.Therefore,it is necessary to optimize the bus scheduling strategy and scientifically apply the dynamic scheduling of bus to improve the service level of the bus system and the utilization rate of bus resources.Based on this,this paper analyzes the dynamic scheduling of single-line buses:(1)Based on the current problems in bus scheduling,such as long passenger travel time and too frequent stops at line stops,a dynamic bus scheduling strategy for the purpose of improving passengers’ travel time satisfaction and bus operation efficiency is proposed.In order to effectively improve the operation efficiency of bus crossing dispatch,the judgment conditions and steps of bus line implementation of cross-station scheduling are proposed,and the research methods of short-term bus passenger flow prediction are analyzed and compared,and the research scheme of adjusting the departure interval of the time period through shortterm bus passenger flow prediction is determined.(2)Through the questionnaire on bus travel,the residents’ satisfaction with the current bus scheduling and the acceptance of bus crossing dispatch are counted,so as to analyze the direction that can be improved in bus scheduling,and provide reliable data support for the determination of model parameters later.In addition,the bus passenger flow data is obtained by the installed automatic counting system,the characteristics of the time and space dimension of the line passenger flow data are analyzed,and the LSTM network model is built according to the passenger flow characteristics to predict the line passenger flow,and the accuracy of the model is verified by comparing with the actual passenger flow,and the departure interval of the unit period is adjusted with the predicted passenger flow data to achieve better results with the subsequent bus crossing station scheduling.(3)A single-line bus crossing station scheduling model considering passenger loss is proposed,which takes into account the cost of additional waiting time caused by crossing the station and the penalty cost of bus enterprises due to the loss of passengers due to crossing the station,and designs a multi-objective function that minimizes the waiting time cost of passenger station,the time cost of passengers on board and the total operating cost of bus enterprises,and uses genetic algorithm to solve the model.(4)Taking the bus operation of route 225 in Xi’an as an example,the bus of this line can be judged whether the bus can implement cross-station scheduling,and the short-term prediction of the passenger flow of the line is made on the basis of the judgment,and the effectiveness of the constructed bus over-station scheduling model considering passenger loss is verified.The example proves that the cost of passenger waiting time,the cost of on-board time and the operating cost of bus enterprises are reduced compared with the dispatch of bus crossing stations.