Multi-agent based model of DDoS active defense

Traditional agent-based modeling is inaccurate and not standard in building network active defense model. To solve such problems, this paper gives the model process of DDoS active defense based on multi-agent. The process models the system from static and dynamic aspects. At the same time, the paper proposes “Goal-Rule-Learn” hybrid agent structure model and “Autonomous-Cooperation-Fight” agent behavioral model. The using of Backus Normal Form description of the agent model gives one standard method of modeling network active defense system. In the end, this paper uses this method builds the investor solid model. The paper makes simulation experiments using OMNeT++. The simulation experiment results verify the correctness and effectiveness of the method.