| The humanization design of traffic simulation system needs the fusion of traffic sociology and traffic engineering, as well as fully integrated subject research activities and achievements. Dynamic traffic scenario integrated simulation system based on Multi-Agent is based on the traditional traffic engineering with traffic behavior theory and the application of artificial intelligence technology. In order to design and realize the system with intelligent, dynamic and integrated multi-level integrated modeling.System is designed to solve the problem, in the complex system, of participants in the intelligence and diversity. For the limitation of traditional simulation system that based on static rules of cellular automata and Multi-Agent, the thought of intelligent query and dynamic integrate the features and rules on participants that based on the knowledge base building by ontology proposed innovative.(1) Put forward the thought of intelligent integrated system for dynamic query and traffic scenarios and put forward the integrated framework of system which is divided into infrastructure layer, mapping layer, rule layer and simulation layer through the analysis of traditional traffic simulation systems and study on Agent theory, cellular automata model and knowledge base.(2) Defined ontology knowledge consists of traffic engineering, traffic behavior and environment domain ontology that based on the cellular automata model and the idea of Multi-Agent, and then set up the base layer using domain ontology, ontology concepts and relations between entities, entities and relationships for data objects.(3) Design method of intelligent query and dynamic integration. Completed the data object intelligent query functions and text rules conversion based on base layer which is dynamic integrated into cellular Agent and participants Agent.(4) Design the Multi-Agent on simulation layer and then take the scenarios simulation through factor decomposition of multi-level integrated modeling that based on the ideal model, participants characteristics and Multi-Agent cooperation model. |
