摘要
为了使入侵容忍系统和可疑者进行最优的策略选择,提出了基于不完全信息动态博弈的入侵容忍技术。根据博弈双方的特点,引入两人随机博弈模型进行分析,并通过精炼贝叶斯纳什均衡来确定双方选择最优的策略。同时为了能够确定可疑者类型,采用熵权和灰色关联分析的方法来进行推理。结合实例使用各种行为特征值来计算灰色关联度,从而在博弈过程中判断可疑者是黑客的概率,验证了该方法的可行性和客观性,从而更好地帮助入侵容忍系统进行策略选择。
In order to help the intrusion tolerance system and suspicious persons to choose the optimal strategies,the intrusion tolerant technology based on incomplete information dynamic game is put forward.According to the characteristics of two sides in the game,the two-player stochastic game models is introduced to analyze the relations,and the perfect Bayesian Nash equilibrium is presented to identify the optimal strategy for them.To estimate the types of suspicious persons,the entropy weight and the gray relational analysis are adopted.With examples,the value of gray correlation degree is calculated by virtue of the behavioral features,and that can help to determine the probability of hackers in the game process,as well to find the optimal strategies for intrusion tolerance system,its feasibility and objectivity are proved.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第11期2430-2431,2486,共3页
Computer Engineering and Design
基金
国家"211工程"重点建设基金项目(181070H901)
关键词
不完全信息动态博弈
入侵容忍
黑客
熵权
灰色关联分析
dynamic game of incomplete information
intrusion tolerance
hacker
entropy weight
gray relational analysis
作者简介
作者简介：娄燕强（1984--），男，安徽合肥人，硕士研究生，研究方向为网络安全、博弈论：
宋如顺，教授，研究方向为网络安全；
马永彩，硕士研究生，研究方向为信息安全。E-mail：louyanqiang@sina．com
