基于强化学习的认知无线网络病毒对抗策略  

VIRUS COUNTERMEASURES IN COGNITIVE WIRELESS NETWORKS BASED ON REINFORCEMENT LEARNING

作　　者：毕凯峰 王健 Bi Kaifeng;Wang Jian(China Southern Power Grid Digital Power Grid Research Institute Co.,Ltd.,Guangzhou 510000,Guangdong,China)

机构地区：[1]南方电网数字电网研究院有限公司,广东广州510000

出　　处：《计算机应用与软件》2025年第2期387-390,共4页Computer Applications and Software

基　　金：南方电网公司科技项目(ZBKJXM20190077)。

摘　　要：认知无线网络通过允许次用户访问主用户拥有的授权频带,显著提高了无线电频谱的利用率。认知无线网络存在被病毒入侵主用户占用授权频带进而降低次用户通信量的可能。提出一种基于多智能体强化学习的认知无线网络病毒对抗策略,通过优化次用户动态路由策略,把病毒攻击对次用户造成的影响降到最低。Cognitive wireless networks improve the utilization of the radio spectrum by allowing secondary users(SU)to access licensed bands owned by primary users.Cognitive wireless network has the possibility that viruses invade primary users and occupy the licensed frequency band to reduce the traffic of secondary users.This paper proposes a cognitive wireless network virus countermeasure strategy based on multi-agent reinforcement learning,which minimizes the impact of virus attack on SUs by optimizing the dynamic routing strategy of SUs.

关 键 词：无线认知网络 强化学习 多智能体强化学习 病毒对抗 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术] TP29[电子电信—通信与信息系统] TN92[电子电信—信息与通信工程]