Design Of Multi-agent Security Model Based On DFL

Aiming at the research condition and the difficulty of agent security research at present, in this paper, we constructed a new agent type-immune agent(IA) by introducing involved biologic immune mechanisms and using Dynamic Fuzzy Logic(DFL) as formal description tool. And then, according to this new kind of agent, we set up a multi-agent security model based on DFL.The main works of this paper are presented as follows:(1) Analyzed the research condition and the difficulty of multi-agent security research at present, and we gave the definition of security in a multi-agent system;(2) Analyzed and sampled involved immune mechanism, we constructed a immune agent and described its mental attribute by DFL;(3) Gave a communication form based on mental attribute;(4) Based on DFL, we proposed a diseriminating model among TIAs by taking into account the theories of dynamic immune network ;(5) Based on the principle of clone selection, we gave a TIA's immune algorithm, and put forward single TIA's distinghuishing model and multi-TIA's distinghuishing model by taking this algorithm as foundation;(6) The security model carried on an application in the detection of computer virus, for this purpose, we also developed a simple simulation program based on binary coding.Through the work above, we gave a new solution for the research of the security mechnisam of Multi-agent and proposed a new method for implementing a feasible multi-agent security model in practice.