Agent modeling and time synchronization method based on multi-agent system

Multi-Agent System is based on agent modeling, distributed interactive, discrete event-driven methods of simulation. It's different from the process-based and time-driven simulation method. Description of agent is hot and difficult research. With Cougaar architecture, according to definition of mixed agent, a general method of modeling Multi-Agent was proposed. It simplified the descriptions of agent model and defined the relationship between any two agents. It also proposed a high performance collaborative algorithm which divided agent into two categories called "master agent" and "simulation agent". It implemented time synchronization with master control, and the agent joined in synchronization when it was necessary. It is possible to reduce the hardware resource consumption effectively and implements effective time synchronization. In addition, each agent can increase or reduce the simulation step adaptively, so it can archive a well balance between simulation performance and accuracy which can target a speciic problem by doing simulation experiments.