Multi-resolution real-time prediction research on tourists’ distribution: a case study of Jiuzhaigou

Tourist’s distribution is a vector in which the ith element refers to the tourist’s arrive number of the node i at a certain time in scenic. A better forecast of tourist’s distribution in real time will produce a better the scenic planning. In this paper, firstly we set up a dynamic recursion prediction model and transform its parameters estimation into optimal control problem. Secondly, time-varying based adaptive dynamic programming algorithm is designed as solution method. Thirdly, a route split method is introduced to split tourist’s sightseeing route into multi-resolution network so that the prediction results can be handled to realize the prediction of multi-resolution in real-time. Finally, Jiuzhai Valley scenic is taken to show the effectiveness of the model.