On Artificial Transportation Systems Modeling And Evaluation Based On Agent

Traffic simulation is an important method to study traffic problems of city, then to improve control and manage efficiency. The concept of artificial transport system provides us with a new perspective to study urban transportation and drives traffic simulation to a higher level.In order to improve the effect of analyzing traffic problems through simulation, the relevant studies of domestic and abroad are analyzed. The thesis proposes the artificial transportation system modeling method of Agent-based accordingly. The potential uses of Agent-based modeling, object-oriented programming, and parallel&distributed computing are investigated to grow artificial transportation system of a typical traffic scene, then we verify the system through traffic flow research data, and discuss the feasibility of artificial transportation systems from modeling practice. Following contents are researched in this thesis:(1) The description of artificial transportation concept, Agent-based modeling techniques and object-oriented programming are executed, then we build the artificial transportation system, verify the feasibility, and establish the study lines of the thesis.(2) Using Agent modeling method to describe adaptive intelligent entity in traffic system in the advantages of its autonomy, sociality, reactivity and interactivity, combining a variety of different functions of Agents, and through interaction, coordination between them, we finally establish the artificial system based on Multi-Agent System, to study the problems of traffic system.(3) Vehicle following model, lane changing model, and signal model in traffic engineering are encapsulated in the Agent entity based on rule. Through the simulation experiment to verify it, the thesis make it to the animation display, and analyze the results by dynamic chart.(4) A large number of traffic flow data survey is carried, real scene traffic investigation data and artificial transportation system simulation data are analyzed by statistics, then we assess the reliability of simulation system by evaluation tree from the perspective of traffic flow, and finally signal optimization experiment is conducted.