【作者】 董晓慧； 陈月辉； 杨波；

【机构】 济南大学信息科学与工程学院；

【摘要】 <正> 1 引言随着网络连接的迅速扩展,特别是Internet大范围的开放,越来越多的系统遭到入侵攻击的威胁。因而,保障计算机系统、网络系统及整个信息基础设施的安全已成为刻不容缓的问题。除了传统的防火墙隔离技术外,安全领域内另一项重要技术和研究方向是入侵检测。目前的网络入侵检测系统多是采用基于规则的检测技术,但是这种技术存在着规则库难于管理、统计模型难以建立及较高的误报率和漏报率等诸多问题,制约了入侵检测系统在实际应更多还原

【Abstract】 A method of detecting intrusion based Evolutionary Neural Network was proposed, which combined Neural Network and Particle Swarm Optimization (PSO). Firstly we used PSO to optimize the weight of the Neural Network. Secondly, we used the optimized Neural Network to detect the intrusion. Evolutionary Neural Network was continuously evolved to search a better Neural Network. When the Evolutionary Neural Network learned the system’s normal working mode and could find the affairs that deviated from the system’s normal working mode, it will find some new attack mode.更多还原