网络系统的容错控制、优化与博弈研究综述  

Review on Fault-Tolerant Control,Optimization,and Game for Network Systems

作　　者：姜斌[1] 许宇航 杨浩[1] JIANG Bin;XU Yuhang;YANG Hao(College of Automation Engineering,Nanjing University of Aeronautics&Astronautics,Nanjing 211106,China)

机构地区：[1]南京航空航天大学自动化学院,南京211106

出　　处：《南京航空航天大学学报》2022年第5期789-800,共12页Journal of Nanjing University of Aeronautics & Astronautics

基　　金：国家自然科学基金(62073165,62020106003,62233009);111引智计划(B20007)。

摘　　要：网络系统是一类由多个子系统通过机械或通信相互耦合所构成的系统。该系统由于其结构复杂、耦合机制多变,具有较高发生故障的概率。随着近几年人工智能的快速发展,网络系统除了可能发生物理故障,还可能存在恶意决策,为系统的安全性带来了新的威胁。本文首先针对网络系统的物理故障,从容错控制和容错优化两个角度总结和梳理当前国内外相关的研究成果。接着,沿着容错博弈控制技术发展的脉络,从博弈控制到分别面向物理故障和恶意决策的容错博弈控制,总结了相关的研究成果。进一步,梳理了当前博弈论在集群飞行器中的应用现状,并以此抛砖引玉,希望推动容错博弈成果在航空航天领域中的应用。最后给出了几个未来值得探索的研究方向。A network system is composed of a couple of subsystems connected with mechanisms or communication.Such a system has a high probability of faults because of its complex structure and changeable coupling mechanisms.With the rapid development of artificial intelligence technology in recent years,network systems are threatened by physical faults and malicious decisions.This paper summarizes current research results on fault-tolerant control and fault-tolerant optimization.Then,following the development of the fault-tolerant game,the relevant research results are summarized from the game control to the faulttolerant game for physical faults and malicious agents.Further,the applications of the game theory on unmanned aerial vehicle(UAV)swarms are summarized.Finally,several research directions that are worth investigating in the future are presented.This review could promote the development of fault-tolerant game in the field of aerospace.

关 键 词：容错控制 博弈控制 互联系统 多智能体系统 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]