基于智能Agent的分布式入侵检测模型设计

Design of Distributed Intrusion Detection Model Based on Intelligent Agent

针对现有入侵检测产品体系结构的局限性和大规模园区网络中的应用需求,依据通用入侵检测框架CIDF,提出了基于改进BP神经网络智能Agent的分布式入侵检测模型,并完成了一个智能Agent的结构设计、算法改进和仿真实验.结果表明,该智能入侵检测单元NIDA的检测准确率超过91.5％,通过多NIDA的设计、训练与互助,该模型可以有效地进行入侵检测.

Aiming at the limitation of the existing IDSs' (Intrusion Detection Systems)structure and the requirement of the application in massive campus network, according to the Common Intrusion Detection Framework(CIDF), this paper proposed a distributed intrusion detection model based on intelligent Agent of improved BackPropagation(BP) neural network, and accomplish its structure design, algorithm improvement, emluator test. The result expresses that the intrusion detection rate of the Neural Intrusion Detection Agent(NIDA) can exceed 91.5 percent, with the design, train and cooperation of multi -NIDA, The model can proceed to detect intrusion effectively.