Evaluation And Optimization Of Rental Stations Configuration Of Lanzhou Public Bicycle System

With the rapid advance of urbanization,the contradiction between urban development and traffic congestion has become increasingly prominent,and the development of public transport is an effective means to alleviate the contradiction.In the choice of public transportation,the subway and bus will undoubtedly become the first choice.However,both the subway and the bus cannot completely solve the terminal traffic problem,that is,the passengers cannot be delivered directly to the destination.In public transport,public bicycle as a means to solve the terminal traffic problem,provide a convenient way to travel.However,due to unreasonable set of public bicycle rental stations,there often exits phenomenon of “no pile for the bicycle or no bicycle for the citizens in public bicycle stations”,which bring difficulties for the public to choose to public bicycles to reach the destination.Therefore,to improve the efficiency of public bicycle system,and guide the public to use public transport travel,become an urgent problem to be solved.Based on analyzing the operation data of public bicycle rental stations in Lanzhou,an evaluation model of public bicycle rental stations setting and lock pile configuration is proposed,in which the public bicycle system is evaluated by improving the C4.5 classification algorithm.Besides,by constructing the simulation model of the public bicycle system based on AnyLogic according to the system dynamics principle,taking the facility utilization rate and service degree of the rental stations as the evaluation index and making use of the combination method of sensitivity analysis and iteration,the number of assigned bicycle of the rental stations is optimized efficiently.The research contents of this thesis are as follows:(1)Statistical analysis of the operation data of the public bicycle system.Based on the actual operation data of the public bicycle system of Lanzhou,the R-language is used to calculate the rental frequency,the return frequency,the number of lock pile and the rental time,and calculate the coupling coordination degree among the data.Traffic terrain characteristics are combined to determine the level of the effect of a single rental station setting.(2)Construct the evaluating decision tree of the effect of public bicycle rental stations setting.The characteristics of these commonly used decision tree classification algorithm are analyzed.Then,the C4.5 classification algorithm was selected,and its calculation process is improved.Taking advantage of the improved C4.5 classification algorithm,the evaluation decision tree for public bicycle rental stations setting and lock-pile configuration is built to obtain the advantages and disadvantages of the rental stations setting.This part of the study provides the basis for the optimization of public bicycle rental stations.(3)Build public bicycle system simulation model and put forward the optimization for the number of assigned bicycle of the rental stations.In view of the low utilization rate of bicycle and the low quality of service,a simulation model of public bicycle system based on AnyLogic is constructed according to the system dynamics principle.Then,the facility utilization rate and service degree of the rental stations are used as the evaluation index,and the number of assigned bicycle of the rental stations is optimized by the combination method of sensitivity analysis and iteration.Finally,the public bicycle system simulation model is used as the experimental platform,and the evaluation index of the rental stations before and after optimization is analyzed and compared.Through the comparison experiment,the results show that the optimized rental station is helpful to improve the efficiency and the service quality of the system.