Study On The Demand Forecast Of Urban Public Bicycle System

With the improvement of the national economy level, motorized travel got rapid development, deterioration of the environment quality, excessive energy consumption,noise pollution and other issues are arised. For these issues, the government put forward the strategy of the "public transport priority development". Public bicycles entered into the limelight as the new part of the public transport system, and it has been taken seriously by more and more people gradually because of the characteristics of environmental and flexible. Public bicycles can solve the problem in the short distance travel and transit traffic demand, and also, the public bicycle system is an important measure to save road resources, reduce environmental pollution and alleviate the difficulty of residents travel.Firstly, the paper introduced the public bicycle system in three aspects, including the system characteristics, elements and function orientation. After that, qualitatively analysed the impact on the public bicycle system that caused by the urban space structure, geographical conditions, traffic facilities and policy factors. Based on the statistical results of questionnaire survey to qualitatively analyse the present use situation of Xi ’an beilin public bicycle system. Then determining the factors that influence the choice behavior of public bicycles to forecast the share ratio of the theory basis of the selection variables of the model. In order to lay a theoretical foundation for the variables selection in the prediction model,the paper identify the factors of public bicycles choice behavior.Secondly, based on a public bicycle travel method,the Bayesian network model of different models were established to calculate the share ratio of public bicycles. BNTtoolbox in MATLAB software was used to learning the structure and parameter of Bayesian network. For the record,using the method of mutual information combined with K2 algorithm to do the structure learning of Bayesian network,using the bayesian estimation to do the parameter learning of Bayesian network. Based on the method,the influences of different factors on the choice of public bicycles were analysed. In order to get the share ratio of the public bicycle system,the model were input into the GENIE software. Calculating the hit ratio of travel modes to test precision of the model.Finally,the paper discussed the calculation of bike and pile’s scale in two aspects,including initial stage and later stage. Using the turnover rate of bikes and the turnover rate of parking piles to calculate the initial scale of the public bicycle system, and established the modified grey prediction model to predict the long-term scale of the public bicycle system. Through modifing the initial condition to improve prediction accuracy. By calculating the relative error to test the prediction accuracy of the model.The model can be used to predict the long-term scale of the public bicycle system because of it’s high precision.