摘要： 应用在计算机集成过程系统(CIPS)网络中的入侵检测系统误报率和漏报率较高。针对该问题，利用遗传算法的全局寻优能力和神经网络对于非线性映射的强大逼近能力，提出具有自学习和自适应能力、基于遗传算法神经网络(GANN)的入侵检测模型，包括数据采集模块、数据预处理模块、神经网络分析模块和入侵报警模块4个部分。为克服遗传算法易早熟、搜索迟钝的缺点，对GANN的适应度值调整方式进行改进，对遗传算法的参数设定进行优化，并采用改进的遗传算法优化收敛速度慢、易陷入极值的BP神经网络。仿真实验结果表明，该模型使系统的检测率提高至97.11%。

关键词: 遗传算法神经网络, BP神经网络, 入侵检测, 计算机集成过程系统, 主成分分析

Abstract: In order to solve the problems of high false alarm rate and fail rate in intrusion detection system of Computer Integrated Process System(CIPS) network, this paper takes advantage that Genetic Algorithm(GA) possesses overall optimization seeking ability and neural network has formidable approaching ability to the non-linear mapping to propose an intrusion detection model based on Genetic Algorithm Neural Network(GANN) with self-learning and adaptive capacity, which includes data collection module, data preprocessing module, neural network analysis module and intrusion alarm module. To overcome the shortcomings that GA is easy to fall into the extreme value and searches slowly, it improves the adjusting method of GANN fitness value and optimizes the parameter settings of GA. The improved GA is used to optimize BP neural network. Simulation results show that the model makes the detection rate of the system enhance to 97.11%.

Key words: Genetic Algorithm Neural Network(GANN), BP neural network, intrusion detection, Computer Integrated Process System(CIPS), Principal Component Analysis(PCA)

中图分类号: 

• TP309.2