Research On Passenger Flow Characteristics And Optimization Of Bus Scheduling Based On Data Mining

Analysis on characteristics of bus passenger flow and optimizing bus scheduling are two important research content of the urban public transport planning and management. With the development of intelligent public transport and the popularity of smart card, more and more cities in China began to use auto fare collection system, but the smart card data is large and comprehensive, has not yet been fully developed. At the same time, the optimization of bus scheduling, which was based on personal experience and lacked the bus passenger distribution characteristics' support, often resulted in uneven transport capacity of public transport vehicles and reduced the attractiveness of public transport system. This paper hopes to use the data mining technology to analyze the passenger flow distribution characteristics, which could be used to support the bus scheduling, fully tap the potential of public transportation, improve the quality of public transport services and bus travel sharing rate of the resident.This paper started with the bus data acquisition technology, focusing on the smart card data acquisition technology and GPS data acquisition technology, describing respectively structure of the bus data. On the basis of the existing research, preprocessing and analysing the data time difference. At the same time this paper obtained the passengers' boarding location and judged the alighting location based on the continuous travel chain theory. One of the innovations of this paper was finding the number of the stations passengers travel obey the Weibull distribution and useing the number of passengers getting on to representative site attraction strength in alighting locations' probability distribution model, and checked the model's error.Secondly, this paper took bus passenger flow distribution characteristics as the support,assuming that the passengers arrive in the uniform distribution,the single line operation interval optimization model was established which took the time of the car waiting time, the crowding time of the travel process and the cost of the bus business as the goal, under constraint condition of the operation interval and the capacity rate, and considered the whole vehicle and interval car two kinds of scheduling form.Finally, this paper optimized bus scheduling of the 16-bus' down direction in Chengdu.We obtained the passengers' boarding/alighting location at each time period. At the same time,the optimal interval was obtained by using the operation interval optimization model, and the corresponding driving timetable was prepared. The validity of the model is verified by comparing with the current scheduling scheme.