Study The Traffic Demand Forecast For Highway Tourism Based On BP Network

The traffic demand forecast of highway tourism is a new research field of both the industry of traffic and tourism .It plays an more and more important role in traffic tourism planning in our days.Firstly, the paper discusses the background, the aim and the signification of the research, and then it summarizes the actualities of the application of ANN in the field of traffic demand forecast. The paper arguments the definition of tourism traffic and classifies the tourism traffic into different modes according to the relationship between tourism and traffic. Some modes of traffic demand forecast of highway tourism which were usually used were explicated and were analyzed in the paper. Then, the paper presents the basic knowledge and the primary network modes of ANN, and analyses ANN's characteristics.Moreover, the paper discusses the principal of traffic forecast of highway tourism and analyses the influence factors of the forecast. Based on the theory of ANN, the lists of parameters to forecast on different mode of tourism traffic were given. Then, the paper explains which network of ANN we choose for the forecast and presents detailedly the BP network and also explains how to apply the BP network to our forecast work.Finally, with an example the paper discusses explicitly the application of BP network in the field of traffic demand forecast of tourism, and then it has a deep research in the ways of the choices of inputting\outputting vector, data pretreatment, the selection on the number of the hidden layer, the selection on the training function, etc. With the analysis of the results, the paper summarizes the advantages and disadvantages of this method and gives some conceivable directions for farther research.