多智能体系统一致性问题中的隐私保护方法综述OA
A survey of privacy-preserving consensus problems in multi-agent systems
多智能体系统(MAS)一致性算法执行中需要个体之间频繁交换并共享信息,这对网络中智能体个体信息的隐私安全带来了巨大的风险.本文阐述了MAS一致性问题中的隐私保护概念定义,给出了相应的问题描述,较系统地回顾了MAS隐私保护一致性技术的现有进展.从方法的归类上将现有技术分为3类:即基于噪声注入的方法、基于密码学的方法和基于状态增强的方法,讨论了针对不同领域实际应用中的设计策略,分析了现有技术的发展趋势,探讨了隐私保护设计中亟待解决的问题与未来的发展方向.
In the execution of consensus algorithms in multi-agent systems(MAS),frequent information exchange and sharing among individuals are required,which poses significant risks to the privacy and security of individual information in the network.This paper introduces the concept definition of privacy-preserving in MAS consensus problems,provides corresponding problem descriptions,comprehensively reviews the existing works of privacy protection technologies for MAS consensus in a systematic manner.Based on the classification of methods,the existing technologies are divided into three categories:methods based on noise injection,methods based on cryptography,and methods based on state decomposition.Design strategies for practical applications in different domains are discussed,and the development trends of existing technologies are analyzed.The unresolved issues in privacy-preserving design and future directions are also explored.
黄明德;伍益明;蒋杰
杭州电子科技大学网络空间安全学院,浙江 杭州 310018杭州电子科技大学网络空间安全学院,浙江 杭州 310018杭州电子科技大学网络空间安全学院,浙江 杭州 310018
多智能体系统协调控制一致性问题隐私保护
multi-agent systemscoordinated controlconsensusprivacy-preserving
《控制理论与应用》 2026 (4)
697-708,12
浙江省公益技术应用研究项目(LGF21F020011),浙江省教育厅科研项目(Y202352122)资助. Supported by the Zhejiang Provincial Public Welfare Research Project of China(LGF21F020011)and the Scientific Research Fund of Zhejiang Provi-ncial Education Department(Y202352122).
评论