基于知识闭环驱动的智能化博弈对抗推演研究

doi:10.16182/j.issn1004731x.joss.25-0685

摘要/Abstract

摘要：

针对人机智能融合和协作增智，提出了“知识-模型-数据-知识”闭环的作战仿真推演范式，指导设计基于DRL的博弈对抗推演架构。构建作战先验知识引导的DRL智能体模型，挖掘分析推演过程产生的智能体交互时序数据，提取形成可扩展指挥人员认知边界的作战后验知识，实现了智能化作战仿真推演的知识闭环驱动机制。实验结果表明：知识闭环驱动机制能有效赋予作战仿真推演系统智能成长能力，为实现作战推演对“人”的认知深化提供了有价值参考。

关键词: 人机智能融合, 博弈对抗推演, DRL, 知识闭环驱动, 智能成长

Abstract:

For human-machine intelligence integration and collaborative intelligence enhancement, a “knowledge-model-data-knowledge” closed-loop paradigm for combat simulation is proposed to guide the design of a DRL-based game confrontation simulation architecture. By building a combat priori knowledge-guided DRL agent model, mining and analyzing the time series data of agent interactions generated during simulations, and extracting combat posterior knowledge that expands the cognition boundaries of commanders, the knowledge closed-loop driving mechanism for intelligent combat simulations is achieved. The experimental results indicate that the proposed mechanism can effectively endow the combat simulation system with intelligence growth capabilities, providing valuable reference for the deepening of “human” cognition in combat simulations.

Key words: human-machine intelligence integration, game confrontation simulation, DRL, knowledge closed-loop driving, intelligence growth

中图分类号:

TP391.9

刘权,王宇,刘林越等 . 基于知识闭环驱动的智能化博弈对抗推演研究[J]. 系统仿真学报, 2026, 38(2): 416-432.

Liu Quan,Wang Yu,Liu Linyue,et al . Knowledge Closed-loop Driving-based Intelligent Game Confrontation Simulation[J]. Journal of System Simulation, 2026, 38(2): 416-432.

图/表 16

图1

图2

图3

图4

图5

表1

表2

主要实验参数

参数	取值
学习率	0.001
动作选择器	$ε$ -greedy
初始探索率	1
最终探索率	0.05
探索率衰减时间步	50 000
并行环境数量	1
批数据量	32
经验回放池容量	5 000
折扣系数	0.99
目标网络更新频率	每200回合
优化器	Adam
回合最大时间步	200

表2

图6

图7

图8

图9

图10

图11

图12

图13

图14

参考文献 28

[1]	胡晓峰. 复杂性也可以成为一种武器[J]. 军事文摘, 2024(1): 1.
[2]	李艺春, 刘泽娇, 洪艺天, 等. 基于多智能体强化学习的博弈综述[J]. 自动化学报, 2025, 51(3): 540-558.
	Li Yichun, Liu Zejiao, Hong Yitian, et al. Multi-agent Reinforcement Learning Based Game: A Survey[J]. Acta Automatica Sinica, 2025, 51(3): 540-558.
[3]	胡晓峰, 齐大伟. 智能化兵棋系统: 下一代需要改变的是什么[J]. 系统仿真学报, 2021, 33(9): 1997-2009.
	Hu Xiaofeng, Qi Dawei. Intelligent Wargaming System: Change Needed by Next Generation Need to be Changed[J]. Journal of System Simulation, 2021, 33(9): 1997-2009.
[4]	胡晓峰. 兵棋推演的"魂"是什么[J]. 军事文摘, 2024(7): 1.
[5]	孙宇祥, 彭益辉, 李斌, 等. 智能博弈综述: 游戏AI对作战推演的启示[J]. 智能科学与技术学报, 2022, 4(2): 157-173.
	Sun Yuxiang, Peng Yihui, Li Bin, et al. Overview of Intelligent Game: Enlightenment of Game AI to Combat Deduction[J]. Chinese Journal of Intelligent Science and Technology, 2022, 4(2): 157-173.
[6]	罗俊仁, 张万鹏, 项凤涛, 等. 智能推演综述:博弈论视角下的战术战役兵棋与战略博弈[J]. 系统仿真学报, 2023, 35(9): 1871-1894.
	Luo Junren, Zhang Wanpeng, Xiang Fengtao, et al. Survey on Intelligent Wargaming: Tactical & Campaign Wargame and Strategic Game from Game-theoretic Perspective[J]. Journal of System Simulation, 2023, 35(9): 1871-1894.
[7]	施伟, 冯旸赫, 程光权, 等. 基于深度强化学习的多机协同空战方法研究[J]. 自动化学报, 2021, 47(7): 1610-1623.
	Shi Wei, Feng Yanghe, Cheng Guangquan, et al. Research on Multi-aircraft Cooperative Air Combat Method Based on Deep Reinforcement Learning[J]. Acta Automatica Sinica, 2021, 47(7): 1610-1623.
[8]	陈浩, 李嘉祥, 黄健, 等. 融合认知行为模型的深度强化学习框架及算法[J]. 控制与决策, 2023, 38(11): 3209-3218.
	Chen Hao, Li Jiaxiang, Huang Jian, et al. Deep Reinforcement Learning Framework and Algorithms Integrated with Cognitive Behavior Models[J]. Control and Decision, 2023, 38(11): 3209-3218.
[9]	韩润海, 陈浩, 刘权, 等. 基于对手动作预测的智能博弈对抗算法[J]. 计算机工程与应用, 2023, 59(7): 190-197.
	Han Runhai, Chen Hao, Liu Quan, et al. Intelligent Game Countermeasures Algorithm Based on Opponent Action Prediction[J]. Computer Engineering and Applications, 2023, 59(7): 190-197.
[10]	Zhou Zhiming, Qi Hongmin, Liu Zhen, et al. Research on Autonomous Decision-making of multi-UAV Air Combat Based on Deep Reinforcement Learning[C]//Proceedings of 2023 7th Chinese Conference on Swarm Intelligence and Cooperative Control. Singapore: Springer Nature Singapore, 2024: 687-700.
[11]	张斌, 雷永林, 李群, 等. 基于强化学习的导弹突防决策建模研究[J]. 系统仿真学报, 2025, 37(3): 763-774.
	Zhang Bin, Lei Yonglin, Li Qun, et al. Reinforcement Learning Modeling of Missile Penetration Decision Based on Combat Simulation[J]. Journal of System Simulation, 2025, 37(3): 763-774.
[12]	胡晓峰, 荣明. 作战决策辅助向何处去-"深绿"计划的启示与思考[J]. 指挥与控制学报, 2016, 2(1): 22-25.
	Hu Xiaofeng, Rong Ming. Where Do Operation Decision Support Systems Go: Inspiration and Thought on Deep Green Plan[J]. Journal of Command and Control, 2016, 2(1): 22-25.
[13]	胡晓峰, 贺筱媛, 陶九阳. AlphaGo的突破与兵棋推演的挑战[J]. 科技导报, 2017, 35(21): 49-60.
	Hu Xiaofeng, He Xiaoyuan, Tao Jiuyang. AlphaGo's Breakthrough and Challenges of Wargaming[J]. Science & Technology Review, 2017, 35(21): 49-60.
[14]	胡晓峰, 荣明. 智能化作战研究值得关注的几个问题[J]. 指挥与控制学报, 2018, 4(3): 195-200.
	Hu Xiaofeng, Rong Ming. Several Important Questions of Intelligent Warfare Research[J]. Journal of Command and Control, 2018, 4(3): 195-200.
[15]	胡晓峰, 齐大伟. 智能决策问题探讨—从游戏博弈到作战指挥, 距离还有多远[J]. 指挥与控制学报, 2020, 6(4): 356-363.
	Hu Xiaofeng, Qi Dawei. On Problems of Intelligent Decision-making—How Far is It from Game-playing to Operational Command[J]. Journal of Command and Control, 2020, 6(4): 356-363.
[16]	贺筱媛, 郭圣明, 吴琳, 等. 面向智能化兵棋的认知行为建模方法研究[J]. 系统仿真学报, 2021, 33(9): 2037-2047.
	He Xiaoyuan, Guo Shengming, Wu Lin, et al. Modeling Research of Cognition Behavior for Intelligent Wargaming[J]. Journal of System Simulation, 2021, 33(9): 2037-2047.
[17]	吴琳, 胡晓峰, 陶九阳, 等. 面向智能成长的兵棋推演生态系统[J]. 系统仿真学报, 2021, 33(9): 2048-2058.
	Wu Lin, Hu Xiaofeng, Tao Jiuyang, et al. Wargaming Eco-system for Intelligence Growing[J]. Journal of System Simulation, 2021, 33(9): 2048-2058.
[18]	Davis P K, Bracken P. Artificial Intelligence for Wargaming and Modeling[J]. The Journal of Defense Modeling and Simulation: Applications, Methodology, Technology, 2025, 22(1): 25-40.
[19]	胡晓峰, 贺筱媛, 陶九阳. 认知仿真:是复杂系统建模的新途径吗?[J]. 科技导报, 2018, 36(12): 46-54.
	Hu Xiaofeng, He Xiaoyuan, Tao Jiuyang. Cognitive Simulation: Is It a New Approach for Complex System Modeling?[J]. Science & Technology Review, 2018, 36(12): 46-54.
[20]	尹奇跃, 赵美静, 倪晚成, 等. 兵棋推演的智能决策技术与挑战[J]. 自动化学报, 2023, 49(5): 913-928.
	Yin Qiyue, Zhao Meijing, Ni Wancheng, et al. Intelligent Decision Making Technology and Challenge of Wargame[J]. Acta Automatica Sinica, 2023, 49(5): 913-928.
[21]	刘潇, 刘书洋, 庄韫恺, 等. 强化学习可解释性基础问题探索和方法综述[J]. 软件学报, 2023, 34(5): 2300-2316.
	Liu Xiao, Liu Shuyang, Zhuang Yunkai, et al. Explainable Reinforcement Learning: Basic Problems Exploration and Method Survey[J]. Journal of Software, 2023, 34(5): 2300-2316.
[22]	Liu Yuntao, Li Yuan, Xu Xinhai, et al. Heterogeneous Skill Learning for Multi-agent Tasks[C]//Proceedings of the 36th International Conference on Neural Information Processing Systems. Red Hook: Curran Associates Inc., 2022: 37011-37023.
[23]	Wang Peng, Wang Haixun, Wang Wei. Finding Semantics in Time Series[C]//Proceedings of the 2011 ACM SIGMOD International Conference on Management of Data. New York: Association for Computing Machinery, 2011: 385-396.
[24]	Zhao Zhe, Wang Pengkun, Wen Haibin, et al. STEM-LTS: Integrating Semantic-temporal Dynamics in LLM-driven Time Series Analysis[C]//Proceedings of the Thirty-Ninth AAAI Conference on Artificial Intelligence and Thirty-Seventh Conference on Innovative Applications of Artificial Intelligence and Fifteenth Symposium on Educational Advances in Artificial Intelligence. Palo Alto: AAAI Press, 2025: 22858-22866.
[25]	Rashid T, Samvelyan M, Christian Schroeder De Witt, et al. Monotonic Value Function Factorisation for Deep Multi-agent Reinforcement Learning[J]. The Journal of Machine Learning Research, 2020, 21(1): 7234 - 7284.
[26]	刘权, 王宇, 付可, 等. 面向城市场景的多智能体协同博弈对抗决策仿真技术研究[C]//2024中国自动化大会论文集. 北京: 中国自动化学会, 2024: 647-652.
	Liu Quan, Wang Yu, Fu Ke, et al. Research on Simulation Technology of Multi-agent Cooperative Game Confrontation Decision-making Simulation for Urban Scenarios[C]//Proceedings of the 2024 China Automation Congress. Beijing: Chinese Association of Automation, 2024: 647-652.
[27]	Cao Yuji, Zhao Huan, Cheng Yuheng, et al. Survey on Large Language Model-enhanced Reinforcement Learning: Concept, Taxonomy, and Methods[J]. IEEE Transactions on Neural Networks and Learning Systems, 2025, 36(6): 9737-9757.
[28]	Liang Yuxuan, Wen Haomin, Nie Yuqi, et al. Foundation Models for Time Series Analysis: A Tutorial and Survey[C]//Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining. New York: Association for Computing Machinery, 2024: 6555-6565.

编组		编组组成
红方	0	2名单兵、1辆无人车、1架无人机
	1	2名单兵、1辆无人车
	2	2名单兵、1辆无人车
蓝方	0	2名单兵、1辆无人车
	1	2名单兵、1辆无人车
	2	3名单兵、1辆无人车