Research On Combined Game Model Of Travel Demand Prediction

Travel demand prediction process is the basis of traffic planning,and the results will be directly used as an analysis basis for the decision makers.In the traditional four-step travel demand prediction method,prediction stages are independent of each other.The prediction work is heavy and the precision is low.By comparison,the combined model can incorporate those closely related stages into a single process,shortening the prediction process and improving the prediction accuracy.Most of the combined models proposed at present only consider the route choice game,but rarely use game theory to study the other stages.In fact,there is crowding effect in choosing whether to travel,departure time,travel destination,travel mode and travel route,forming a multi-dimensional game of group travel.However,it is still rare to establish a multi-step combination model based on game theory.Therefore,based on the potential game theory,a multi-step combined model is established in this paper.The main research contents are as follows:1.Two-step combined game model of travel destination and travel route choiceFirstly,the application of the potential game theory in travel destination and travel route choice behavior modeling is introduced.On this basis,we study the two-step combined game model of travel destination and travel route choice,analyze the equilibrium conditions and propose the mathematical programming model.Finally,a combined model solving algorithm is designed based on the network transformation and an example is given.2.Three-step combined game model of travel destination,travel mode and travel route choiceThis study firstly defines a travel mode choice game,uses the super-network theory to construct a network with multiple travel modes and further gives the cost function of various road sections.Then we analyze the equilibrium conditions,propose the mathematical programming model and design a solving algorithm.Finally,a three-step combined game model of travel destination,travel mode and travel route choice is established.3.Practical application of combined game modelTaking the city of Cangzhou as the application object and relying on the data of WordPop and urban road network,the traffic communities are divided and the population of each traffic community is extracted with ArcGIS.And then the amount of traffic volume at the early peak is calculated.With the traffic volume data,the road network data and the combined game model,we obtain the travel distribution between traffic cells,the travel attraction of each cell and the travel flow on the network under equilibrium condition.By plotting the flow Saturation diagram,the congestion level and service level of each section is intuitively reflected.