Research On Bi-level Planning Of Demand Responsive Transit Lines Considering Dynamic Stations

With the rapid development of China’s social economy after a long period of time,the city’s outward expansion and changes in the characteristics of residents’ travel needs have provided new challenges for urban transportation.The number of private cars is on the rise every year,and urban traffic congestion and traffic pollution are getting worse.Under such circumstances,the energy saving and environmental protection of public transportation and its intensive efficiency provide an effective means for solving urban transportation problems.However,with the decentralization of urban centers,the density of residents’ travel in most areas and the lack of aggregation,traditional public transportation services cannot meet the travel needs of different types of people.Demand-responsive public transport,as a form of public transportation that has been in development for some time,combines the advantages of high efficiency and low cost of fixed-line public transportation and the convenience and flexibility of connecting public transportation.At the same time,the advanced Internet technology and communication transmission speed in the era of "Internet +" and "big data" also provide technical support for various types of demand response bus modes.Therefore,this article took the demand-responsive bus system and passenger travel behavior selection as the research objects,and focused on the demand-responsive bus system considering dynamic stations.First,the related concepts of demand-responsive buses considering dynamic stations were described,and then on the basis of two bus operation modes of fixed-line buses and connected demand-responsive buses,a demand-responsive bus system model considered dynamic stations.To provide a reference for whether the model’s demand-responsive public transport was suitable for China’s urban and rural environment.Then,for this bi-level programming problem,a four-stage solution method was designed.The first stage:used K-means clustering method to determine fixed stations,and then used DBSCAN clustering algorithm to determine the service scope of all demand response stations.The second stage:first applied the improved simulated annealing algorithm to solve the upper model to solve the dynamic demand response site activation problem.The third stage:based on the first part of the upper objective function,the prediction of the company’s revenue,and then used the accurate solution to solve the lower model,got the passenger’s response to the activated station,and determined the number of passengers who will use this system.The fourth stage:through the scheme determined in step three,assigned the activated dynamic station to each bus,determined its driving route,and used the GASA algorithm to solve the upper model problem.Through MATLAB programming and Cplex solver to implement the algorithm of the model,and completed the applicability and sensitivity analysis of the model.Finally,selected the travel survey data of residents in Lufeng City,combined with their station information,passenger flow information,vehicle configuration,vehicle operation information and other data,to carry out an example analysis of the dynamic station demand response type bus line bi-level planning model.