中文摘要：

针对当前基于主机的入侵防御系统的实时性、自适应性已经不能满足计算机网络安全需求的问题,提出一个具备认知功能的入侵防御系统方案.该方案由传感器、状态库、知识库、认知推理和决策执行等部分组成,传感器将感知到的环境信息经过知识库、状态库和认知推理形成的认知环的处理之后,产生一个应对入侵的决策方案,利用优先级库确定自适应尝试的启发式搜索方法,形成一个反馈循环来完成系统的自学习.该文对模型的每一部分的功能及实现机理进行了分析描述.

英文摘要：

Because of the instantaneity and self-adaption of present intrusion prevention system （IPS） can not fulfill the current security needs of computer and network systems,an IPS model which has the cognitive ability is proposed.This program is consisted of sensors,state library,konwledge base,cognitive inference and decision-execution.First of all,a decision-execution mechanism is produced when the environment information perceived by the sensors goes through a cognitive cycle which formed by the knowledge base,state library and cognitive inference.Then,an adaptive heuristic searching method which is confirmed by the priority library is coordinated,a feedback loop is established.Finally,the self-learning process is completed.Also the function and the mechanism of each part of this model is introduced.