摘要
针对实战中红蓝双方对抗博弈过程中的决策效率低、决策规则简单、智能化程度低等问题,采用人工智能与传统仿真平台相结合的方法,构建智能自博弈平台,训练出具有决策能力的红方智能体和蓝方智能体。通过1V1空战仿真试验表证明,红蓝双方智能体具备决策能力,可以生成较好水平的交战决策,能够完成智能自博弈的过程。试验结果和研究内容对推进未来军事仿真智能化具有参考意义。
In order to solve the problems of low decision-making efficiency,simple decision rules,and low intelligence during the game between red-party and blue-party on the battlefield,an intelligent selfgame platform is constructed,by using the combination of artificial intelligence and traditional simulation platform.In the intelligent self-game platform,the red-party agent and the blue-party agent with decisionmaking ability can be trained and made intelligent in the platform.Through the 1 V1 air combat simulation experiment,it is proved that the red and blue agents in the experiment which have the decision-making ability,can generate a good level of engagement decision,and can complete the process of intelligent selfgame.The results of experimental and research have guiding significance for advancing the future military simulation intelligence.
作者
卢锐轩
孙莹
杨奇
王壮
吴昭欣
李辉
Lu Ruixuan;Sun Ying;Yang Qi;Wang Zhuang;Wu Zhaoxin;Li Hui(College of Computer Science,Sichuan University,Chengdu 610065,China;Unit 31001 of the PLA,Beijing 101400,China)
出处
《战术导弹技术》
北大核心
2019年第2期47-52,98,共7页
Tactical Missile Technology
关键词
自博弈
人工智能
智能体
军事仿真智能化
self-game
artificial intelligence
agent
military simulation intelligence
