首页 | 本学科首页   官方微博 | 高级检索  
     

多Agent系统安全保护机制研究
作者姓名:范若楠  张晓荣  王鑫  李斌  张俊
作者单位:中国船舶重工集团公司第七〇五研究所,陕西西安, 710000;西安交通大学计算机科学与技术学院,陕西西安, 710049
摘    要:针对现有多Agent系统易受到恶意攻击、难以验证代理身份及代理间难以建立信任机制等安全问题,围绕多Agent分层系统架构,提出一种采用基于数字签名和非对称加密的多Agent通信加密方法,有效保证了通信环节的系统安全,防范了Agent代理身份被冒用。基于动态博弈对抗的思路,在多Agent体系评估的基础上,利用贪婪算法和遗传算法优化多Agent系统安全选择策略,提高了防御者在攻防博弈中获胜概率并缩短了攻防博弈轮次,保证了通信环节的安全防御,防止Agent代理信息被攻击者窃取。

关 键 词:多Agent系统  加密通信  数字签名  贪婪算法  遗传算法

Research on multi-agent system security protection mechanism
Authors:FAN Ruonan  ZHANG Xiaorong  WANG Xin  LI Bin  ZHANG Jun
Abstract:Aiming at the security problems such as the existing multi-agent system is vulnerable to malicious attacks, it is difficult to verify the identity of the agent and to establish a trust mechanism between agents, this paper focuses on the multi-agent layered system architecture, and proposes a multi-agent encryption based on asymmetric encryption algorithm and digital signature. The communication method can effectively ensure the system security of the communication and prevent the agent identity from being used fraudulently. Based on the dynamic game ideas in multiple agent system assessment, using the greedy algorithm and genetic algorithm to optimize the multi-agent system security selection strategy, we effectively increase the probability of winning and shorten the rounds of gaming, ensure the security of the communication, and prevent the agent information from being eavesdropped.
Keywords:
点击此处可从《》浏览原始摘要信息
点击此处可从《》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号