火力与指挥控制2023，Vol.48Issue(12)：8-16,26,10.DOI:10.3969/j.issn.1002-0640.2023.12.002
强化学习在军事上的应用
An Overview on the Military Applications of Reinforcement Learning
宋泠澳 1刘涛 1赵冬梅 1董宏扬1
作者信息
- 1. 西南科技大学计算机科学与技术学院,四川 绵阳 621010
- 折叠
摘要
Abstract
As a hot topic in the field of artificial intelligence,reinforcement learning is widely used in many fields,including the military field.The development of reinforcement learning and the basic theory of reinforcement learning are expounded;four kinds of classical approaches to of reinforcement learning are introduced:According to such reinforcement learning based on strategy,value,model,and model-free and the cutting-edge researches on reinforcement learning in military fields are overviewed,including in unmanned aircraft operations,missile target situational assessment,robot control,path planning of underwater unmanned vehicles,etc.The research shows that reinforcement learning can provide fast decision-making capability for military applications,especially empowering unmanned equipment such as missiles,UAVs,and robots,etc.with strong autonomous decision-making capabilities,thereby improving their destructive or defensive capabilities and enhancing the intelligentialization level of the equipment.Meanwhile,reinforcement learning has been preliminarily applied in complex military applications such as target assignment and defense,complex data fusion,dynamic collaboration for supply assurance,and deck scheduling problems for aircraft carriers to provide fast and accurate decision support.The challenges and future developments of reinforcement learning in military applications are summarized finally.关键词
强化学习/军事应用/算法/智能体Key words
reinforcement learning/military applications/algorithms/agent分类
信息技术与安全科学引用本文复制引用
宋泠澳,刘涛,赵冬梅,董宏扬..强化学习在军事上的应用[J].火力与指挥控制,2023,48(12):8-16,26,10.
