Passenger Flow Prediction And Scheduling Optimization Of Urban Public Transportation In Rainy Weather

It is one of the effective means to alleviate urban traffic congestion that giving priority to the development of urban public transport.Especially,under rainy weather conditions,it is one of the urgent problems to improve the utilization efficiency of public transportation resources and increase the attractiveness of urban public transportation.This paper explores the fluctuation law of bus passenger flow affected by rainfall,and puts forward a passenger flow prediction model under rainy weather on the basis of the influencing factors of bus passenger flow.The departure intervals of buses under different levels of rainfall are optimized on the basis of the prediction model.Firstly,the bus data and rainfall data are preprocessed,at the same time,the changing characteristics of passenger flow are analyzed.Then the influence degree of rainy weather on bus passenger flow from the aspects of day type,time period and season are analyzed by this paper.At last,the fluctuation characteristics of bus passenger flow in rainy weather are obtained.Secondly,the SVR(Support Vactor Regression)model and LSTM(Long Short-Term Memory)are selected to forecast the hourly passenger flow of Bus No.45 in Xiamen on the basis of the influence law of rainfall factors on bus passenger flow.Then the LSTM model is modified to construct a bus passenger flow prediction model in rainy weather on the basis of the analysis of the influence of rainfall on time-sharing passenger flow.Finally,the effectiveness of the modified model is verified by comparing the prediction results of different models.Lastly,a city bus dispatch model is constructed based on the predicted bus passenger flow under rainy weather conditions,with the objective function of minimizing the total social cost considered from the perspective of bus enterprise operation and passenger costs,and with vehicle load rate and departure interval constraints as conditions.The departure time intervals under different rainfall weather conditions are optimized using the NSGA-II(non-dominated sorting genetic algorithm-II).A prediction model of bus passenger flow in rainy weather is proposed on the basis of the changing law of bus passenger flow in rainy weather,and the departure time intervals of buses are optimized in this paper.This provides a basis for the bus operation department to respond to bus scheduling and organization under rainy weather conditions,and has certain theoretical significance and practical value.