Uncertain Transit Demand Forecast Based On IC Card And GPS Data

With the contrast in development of city scale and its road system,this supply-and-demand dilemma causes severe deterioration of urban traffic,restricting the sustainable development of city and making public transportation the trend in the future.As an important part of the planning of public transportation system,demand forecasting of public transportation has always been the focus of the research for scholars at home and abroad.On the basis of the regular bus data of Shenzhen city,and combining with the Interval Uncertainty Theory,the author puts forward the research target,the interval uncertainty of demand for bus by IC(Integrated Circuit)card and GPS(Global Positioning System)data,with its focus on bus IC card data and bus stop identification through GPS data,get-off stop inference,the calculation of interval bus demand and the assignment of indefinite interval get-off passenger on the premise of IC card data and GPS data.To begin with,the methods,tools and theories used for bus data processing were analyzed,the date and the existing problems of regular bus of Shenzhen city was expounded in the paper,Secondly,the existing Interval Uncertainty Theory and the calculation methods of interval numbers were analyzed in this thesis.It determined the confidence interval,the quartile interval,the quintile interval,and derived the bus passenger flow interval analysis method based on its lines and areas;Thirdly,the data of Shenzhen regular bus traffic was preprocessed in this thesis,worked out a method to spot out the get-on stops through the joint usage of IC card and GPS data,the results worked out by which were then applied in the Interval Uncertainty Analysis and tested to enable the identification of the best-for-getting-on intervals and a feature analysis on passenger flow of buses;Then,the deducing models and processes for individual lines under the transfer behavior are improved through analyzing the bus card swiping data and integrating the characteristics of individual passengers,their travel distance in stops,and multi-day travel chains.The results by these models were then derived,and an expansion of sample analysis was pursued to acquire the bus lines passenger flow OD(Origin Destination)and conduct the interval uncertainty analysis to find out bus lines interval passenger OD.It then divided the public transport areas to get the bus area interval traffic OD,and ran date tests to find out the best passenger flow interval.Finally,in view of the interval uncertainty demand and interval uncertain impedance and attempting to improve from the two aspects of travel time function and transfer function,an improved Logit model allowing the distribution of uncertain interval bus passenger flows with-and-without transfers was set up.The model was then verified through error analysis to obtain the bus interval passenger flow distribution.This research on bus interval uncertainty demand forecasting,based on bus IC card and GPS data,is to some extent innovative and forward-looking.It optimizes the derivation of the previous alighting stations by combining the individual characteristics of passengers,the multi-day travel chain and the travel distances measure by stations;It for the first time,puts forward the perspective of bus interval demand and provided a planning reference to make public transport planning more tally with the actual situation.Thus,it provides some sort of references for the development of bus demand forecasting and relevant academic researches.