A Public Traffic Demand Forecast Method Based On Computational Experiments

With the rapid development of social economy,the application of travel tools is becoming more and more popular,which has led to increasingly serious urban traffic problems that even become a bottleneck restricting the development of national economy.So the study of traffic demand forecast is of great significance.The most widely used method for the study of traffic demand forecast is the four-stage method,but the traditional mathematical models commonly used in its each step lack the consideration of the heterogeneity of the actual travel individual and the dynamic nature of the traffic environment,so they can not forecast the traffic demand in different traffic scenarios,and their accuracy still need to be improved.In order to overcome the limitations of the traditional mathematical models,this paper proposed a computational experiments approach for public traffic demand forecast.In this paper,the main framework of the computational experiments approach for public traffic demand forecast,which is mainly composed of three parts: traffic survey,artificial transportation system and computational experiments,is introduced in detail.In the traffic survey,in addition to the traditional roadside traffic survey,this paper also presents a kind of individual-oriented traffic survey,in order to further explore the individual travel behavior mode.In the artificial transportation system,this paper adopts the agent-based modeling method to respectively model the three main subjects in transportation system: travel individuals,travel tools and traffic environment.The BDI model is introduced into the modeling of travel individuals to realize the deduction of the heterogeneity travel individuals' traffic decision-making process in different traffic scenarios combined with their own transportation demands,habit preferences and external information.In this paper,computational experiments is divided into two categories according to the experimental purpose: the traffic demand forecast experiments in the specific traffic situation and the computational experiments of discovering the hidden law in the traffic demand.The school bus system in Huazhong University of Science and Technology is studied as a case in this paper.We conducted a number of traffic surveys,and use Anylogic software to build an artificial transportation system.By comparing with the actual traffic survey data in the same study period,the accuracy and feasibility of the proposed public traffic demand forecast method are verified.On this basis,this paper also realizes the forecast of traffic demand under various traffic scenarios,explores the influence level of various influencing factors on the traffic division,and provides decision support for the school bus distribution scheme in the case system.