一种面向IP网络安全的网络智能体模型

Based on the HIS antibody mechanic, a controllable agent model is presented to protect and enhance survivability of current networks from intelligent attacks in this paper. The agent can detect potential danger actively and isolate infected nodes timely. Meanwhile, it also report on signatures to the local B node to achieve information sharing. Integrated negative selection theory and danger theory, a new training method of agent proposed so that the network can more effectively identify the node on attacking. By analyzing the anti-model about agent and worm, the model and simulation results show that the network agent model is able to protect network security better.