Distributed Discrete Event Simulation Platform

With the development of computer technology, as a modern scientific research method, computer stimulation generally have the attention of people. But the stimulation method is becoming utmost complex, thus the computing environment with single-process cannot meet the need of practical research. So people is trying to solve the problems using multi-process, which can be offered by paralleled or distributed environment.The goal of this thesis is developing a universal distributed stimulation platform, based on multi-agent theory, researching from universal and distributed direction. Using this platform, researchers could develop a stimulation program focusing on the stimulation model itself, while not concern much of the distributed computer programming. By simple developing processes, researcher could develop high performance and reliability stimulation application.In this thesis, we focus on time management in the distributed stimulation. After analysis some classical methods in time management, such as optimum or conservation, we point out some weak points of the current ways. Finally, the thesis put up integrated methods.In the meanwhile, the research designs a set of framework based on layered structure, from middleware and multi-agent point. In this process, we redevelop the platform based on an open-source project madkit, adding time management layer.Last but not the least, the thesis direct the further research direction based on the current research.