多智能体系统安全性问题及防御机制综述  

A survey of the security issues and defense mechanisms of multi-agent systems

作　　者：丁俐夫 颜钢锋 DING Lifu;YAN Gangfeng(College of Electrical Engineering,Zhejiang University,Hangzhou 310027,China;Huanan Industrial Technology Research Institute,Zhejiang University,Guangzhou 510760,China)

机构地区：[1]浙江大学电气工程学院,浙江杭州310027 [2]浙江大学华南工业技术研究院,广东广州510760

出　　处：《智能系统学报》2020年第3期425-434,共10页CAAI Transactions on Intelligent Systems

基　　金：国家重点研发计划项目(2018YFB0904900).

摘　　要：多智能体系统作为分布式人工智能的重要分支,已成为解决大型、复杂、分布式及难预测问题的重要手段。在开放网络中,多智能体系统仍面临许多安全问题,潜在的安全威胁很可能影响其实际应用的稳定性、快速性和准确性。基于目前已知的多智能体系统通用模型,介绍了多智能体系统通信协议、访问控制和协调机制中潜在的安全问题,规范了多智能体系统安全性问题的研究体系,总结了系统设计过程中可行的防御技术和隐私保护技术,最后展望了多智能体系统安全研究的发展方向。Currently,multi-agent systems are an important branch of distributed artificial intelligence and an important means of solving large,complex,distributed,and difficult-to-predict problems.However,in an open network,multiagent systems still face many security issues.Potential security threats are likely to affect the stability,speed,and accuracy of the practical applications of multi-agent systems.Based on the general models of currently known multi-agent systems,the potential security problems in multi-agent system communication protocol,access control,and coordination mechanism are introduced.Research on multi-agent system security problems is standardized,and feasible defense and privacy protection technologies in the system design process are summarized.Finally,the development trend of multi-agent system security research is analyzed.

关 键 词：多智能体系统 分布式人工智能 安全威胁 防御机制 网络安全 通信协议 访问控制 协调机制 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP309[自动化与计算机技术—控制科学与工程]