一种信息不确定系统的模糊安全博弈模型

发布时间：2018-07-26 12:53

【摘要】：面对信息技术广泛运用的今天,信息安全问题开始受到人们的广泛关注.现有的最新工作大多使用贝叶斯博弈的方法来解决.贝叶斯博弈可以应对参与者类型信息不完全的情况,但现实生活中,信息不仅不完全而且是不确定的,即不知道对手类型精确的分布概率.为解决这个问题,扩展了贝叶斯博弈,提出极小极大模糊博弈模型(ABGMR).该模型中使用证据理论来描述信息的模糊性;引入可接受收益的极小极大遗憾原则,在收益可接受的情况下最小化最大遗憾值,避免了出现无法接受的收益和过度悲观的情况.极小极大模糊博弈模型适用于多种攻击者多个防御者的情况,并且攻击者和防御者不需先观察对手的策略,可以同时采取行动.最后将ABGMR运用在信息安全领域资源分配场景中,并与已有博弈算法DOBSS进行实验对比,证明了其有效性和最终决策的最大遗憾值的优越性.
[Abstract]:In the face of the widespread use of information technology today, information security issues have been widely concerned. Most of the latest work is solved by the method of Bias game. The Bias game can cope with incomplete information on the types of participants, but in real life, the information is not only incomplete and uncertain, that is, I do not know. In order to solve this problem, to solve this problem, the Bias game is extended and the minimax fuzzy game model (ABGMR) is proposed. The model uses evidence theory to describe the fuzziness of information; the minimax regret principle of acceptable income is introduced to minimize the maximum regrets under acceptable benefits and avoid the maximum regret value. There are unacceptable returns and excessive pessimism. The minimax fuzzy game model is suitable for multiple attackers, and the attacker and defender do not need to observe the opponent's strategy first and take action at the same time. Finally, ABGMR is applied to the resource allocation scene in the information security field, and the game is calculated with the existing game. The comparison of the method DOBSS proves the superiority of the method and the maximum regret value of the final decision.
【作者单位】：南京工业大学计算机科学与技术学院;
【基金】：江苏省自然科学基金项目(BK20150960)资助
【分类号】：O225;TP309

【相似文献】