对抗场景下基于多智能体联盟更新的指控逻辑链优化问题研究

doi:10.12305/j.issn.1001-506X.2026.08.23

系统工程与电子技术 ›› 2026, Vol. 48 ›› Issue (8): 2782-2789.doi: 10.12305/j.issn.1001-506X.2026.08.23

• 系统工程 • 上一篇

对抗场景下基于多智能体联盟更新的指控逻辑链优化问题研究

张杰¹^,²(), 王超¹, 刘玉¹, 李东¹, 陈志群¹, 吕天启¹

1. 中国电子科技集团公司第二十八研究所，江苏南京 210007
2. 南京大学计算机学院，江苏南京 210023

收稿日期:2025-01-07 修回日期:2025-05-30 出版日期:2026-03-20 发布日期:2026-03-20
通讯作者: 张杰 E-mail:guyuexiao95@gmail.com
作者简介:王　超（1988—），男，高级工程师，博士，主要研究方向为指挥信息系统
刘　玉（1985—），男，高级工程师，硕士，主要研究方向为指挥信息系统
李　东（1991—），男，高级工程师，博士，主要研究方向为指挥信息系统
陈志群（1991—），男，工程师，博士，主要研究方向为指挥信息系统
吕天启（1992—），男，工程师，博士，主要研究方向为指挥信息系统

Research on command and control logic chain optimization based on multi-agent coalition updates in adversarial scenarios

Jie ZHANG¹^,²(), Chao WANG¹, Yu LIU¹, Dong LI¹, Zhiqun CHEN¹, Tianqi LU¹

1. The 28th Research Institute of China Electronics Technology Group Corporation，Nanjing 210007，China
2. School of Computer Science，Nanjing University，Nanjing 210023，China

Received:2025-01-07 Revised:2025-05-30 Online:2026-03-20 Published:2026-03-20
Contact: Jie ZHANG E-mail:guyuexiao95@gmail.com

摘要/Abstract

摘要：

对抗场景中，由于地形复杂、环境变化快、战场信息碎片化等因素，如何通过多智能体联盟进行动态合作，实时更新和优化逻辑链的各个环节，是提高作战效能的核心挑战之一。本文将对抗场景逻辑链构建问题建模为多智能体联盟构建问题，针对联盟决策效率低下和合作不稳定的挑战，提出一种基于博弈论和机制设计的联盟更新机制。该机制结合强化学习技术，优化多智能体间的动态博弈行为，从而提升联盟在复杂对抗环境下的合作效率。引入的自适应机制设计，能够根据战场态势的变化动态调整联盟成员的决策规则，确保联盟的灵活性和鲁棒性。实验结果表明，所提方法在多种对抗场景下显著提高了联盟的作战效能和协同表现，验证了其有效性，可为对抗场景作战体系中的智能体联盟构建提供理论支持和技术指导。

关键词: 博弈论, 机制设计, 逻辑链构建, 强化学习

Abstract:

In adversarial scenario, due to factors such as complex terrain, rapid environmental changes, and fragmented battlefield information, how to conduct dynamic cooperation through multi-agent coalition, and update and optimize all links of the logic chain in real time, is one of the core challenges to improve operational effectiveness. This paper models the problem of constructing a land warfare logic chain as a problem of constructing a multi-agent alliance. Aiming at the challenges of low coalition decision-making efficiency and unstable cooperation, a coalition update mechanism based on game theory and mechanism design is proposed. Combined with reinforcement learning technology, the dynamic game behavior among multi-agents is optimized, thereby improving the cooperation efficiency of the coalition in complex confrontation environments. The introduced adaptive mechanism design can dynamically adjust the decision-making rules of coalition members according to changes in battlefield situation, ensuring the flexibility and robustness of the coalition. Experimental results show that the proposed method significantly improves the operational effectiveness and collaboration performance of the coalition in various adversarial scenarios, verifying its effectiveness. It can provide theoretical support and technical guidance for the construction of agent coalition in adversarial scenarios.

Key words: game theory, mechanism design, logic chain construction, reinforcement learning

中图分类号:

TP18
E11

张杰, 王超, 刘玉, 李东, 陈志群, 吕天启. 对抗场景下基于多智能体联盟更新的指控逻辑链优化问题研究[J]. 系统工程与电子技术, 2026, 48(8): 2782-2789.

Jie ZHANG, Chao WANG, Yu LIU, Dong LI, Zhiqun CHEN, Tianqi LU. Research on command and control logic chain optimization based on multi-agent coalition updates in adversarial scenarios[J]. Systems Engineering and Electronics, 2026, 48(8): 2782-2789.

图/表 4

参考文献 34

1	刘祥雨, 王刚, 郭相科, 等. 面向区域防空场景的杀伤链设计方法[J]. 系统工程与电子技术, 2025, 47 (5): 1582- 1599.
	LIU X Y, WANG G, GUO X K, et al. Design method of kill chain for regional air defense scenarios[J]. Systems Engineering and Electronics, 2025, 47 (5): 1582- 1599.
2	钱林方, 张龙, 佟明昊, 等. 面向信息化智能打击的现代火炮技术发展与展望[J]. 陆军工程大学学报, 2025, 4 (2): 1- 8.
	QIAN L F, ZHANG L, TONG M H, et al. Development and prospect of modern artillery technology for information-based intelligent strike[J]. Journal of Army Engineering University, 2025, 4 (2): 1- 8.
3	张昊, 邓大松. 未来无人反无人杀伤链展望分析[J]. 战术导弹技术, 2024, 1 (5): 160- 167.
	ZHANG H, DENG D S. Prospective analysis of future unmanned counter-unmanned kill chain[J]. Tactical Missile Technology, 2024, 1 (5): 160- 167.
4	JOHNSON B, GREEN J M, BURNS G, et al. Mapping artificial intelligence to the naval tactical kill chain[J]. Naval Engineers Journal, 2023, 135 (1): 155- 166.
5	王耀祖, 尚柏林, 宋笔锋, 等. 基于杀伤链的作战体系网络关键节点识别方法[J]. 系统工程与电子技术, 2023, 45 (3): 736- 744.
	WANG Y Z, SHANG B L, SONG B F, et al. Key node identification method of combat system network based on kill chain[J]. Systems Engineering and Electronics, 2023, 45 (3): 736- 744.
6	ZHANG Z L, TIAN S P, KONG L T. Application for over-the-air computation in command and control data link systems[J]. Command Information System and Technology, 2022, 13 (6): 55- 62.
7	XU C T, ZHOU F, CHEN H H, et al. Autonomous adaptive capability evaluation method for intelligent command and control system[J]. Command Information System and Technology, 2023, 14 (6): 29- 38.
8	ZHANG J L, HAN K, ZHANG P, et al. A survey on joint-operation application for unmanned swarm formations under a complex confrontation environment[J]. Journal of Systems Engineering and Electronics, 2023, 34 (6): 1432- 1446. doi: 10.23919/JSEE.2023.000162
9	徐成涛, 周芳, 陈洪辉, 等. 智能化指挥控制系统自主适变能力评估方法[J]. 指挥信息系统与技术, 2023, 14 (6): 29- 38.
	XU C T, ZHOU F, CHEN H H, et al. Assessment method for autonomous adaptive capability of intelligent command and control system[J]. Command Information System and Technology, 2023, 14 (6): 29- 38.
10	张福林, 孔凡照, 黄凯. 基于智能调度策略的火力规划技术[J]. 指挥信息系统与技术, 2024, 15 (4): 46- 49，62.
	ZHANG F L, KONG F Z, HUANG K. Firepower planning technology based on intelligent scheduling strategy[J]. Command Information System and Technology, 2024, 15 (4): 46- 49，62.
11	GAO X L, WANG R, LIU Z H, et al. A land-based war-gaming simulation method based on multi-agent proximal policy optimization[C]//Proc. of the IEEE 24th International Conference on Software Quality, Reliability, and Security Companion, 2024: 611–618.
12	BLACK K, JANNER M, DU Y, et al. Training diffusion models with reinforcement learning [EB/OL]. [2025-07-01]. https://arxiv.org/abs/2305.13301.
13	KORKMAZ E. Adversarial robust deep reinforcement learning requires redefining robustness[C]//Proc. of the AAAI Conference on Artificial Intelligence, 2023: 8369–8377.
14	CHEN Y J, ZHENG Z C, GONG X L. Marnet: backdoor attacks against cooperative multi-agent reinforcement learning[J]. IEEE Trans. on Dependable and Secure Computing, 2022, 20 (5): 4188- 4198.
15	XU Z W, BAI Y P, ZHANG B, et al. Haven: hierarchical cooperative multi-agent reinforcement learning with dual coordination mechanism[C]// Proc. of the AAAI Conference on Artificial Intelligence, 2023: 11735–11743.
16	WU Z F, YU C, CHEN C, et al. Models as agents: optimizing multi-step predictions of interactive local models in model-based multi-agent reinforcement learning[C]// Proc. of the AAAI Conference on Artificial Intelligence, 2023: 10435–10443.
17	BOGGESS K, KRAUS S, FENG L. Explainable multi-agent reinforcement learning for temporal queries[C]// Proc. of the 32nd International Joint Conference on Artificial Intelligence, 2023: 55–63.
18	CHEN L, LU K, RAJESWARAN A, et al. Decision transformer: Reinforcement learning via sequence modeling[J]. Advances in Neural Information Processing Systems, 2021, 34 (1): 15084- 15097.
19	CHRISTIANOS F, PAPOUDAKIS G, RAHMAN M A, et al. Scaling multi-agent reinforcement learning with selective parameter sharing[C]//Proc. of the International Conference on Machine Learning, 2021: 1989–1998.
20	LIU I J, JAIN U, YEH R A, et al. Cooperative exploration for multi-agent deep reinforcement learning[C]//Proc. of the International Conference on Machine Learning, 2021: 6826–6836.
21	文永明, 李博研, 张宁宁, 等. 基于深度强化学习的多智能体编队协同控制[J]. 指挥信息系统与技术, 2023, 14 (5): 75- 79.
	WEN Y M, LI B Y, ZHANG N N, et al. Multi-agent formation cooperative control based on deep reinforcement learning[J]. Command Information System and Technology, 2023, 14 (5): 75- 79.
22	唐志一, 蔡颖, 王会. 基于自适应加权算法的多传感器数据融合方法[J]. 指挥信息系统与技术, 2022, 13 (5): 66- 70.
	TANG Z Y, CAI Y, WANG H. Multi-sensor data fusion method based on adaptive weighted algorithm[J]. Command Information System and Technology, 2022, 13 (5): 66- 70.
23	KOLPACZKI P, BENGS V, MUSCHALIK M, et al. Approximating the shapley value without marginal contributions[C]//Proc. of the AAAI Conference on Artificial Intelligence, 2024: 13246–13255.
24	LIU L, HU Y J, GAO Y, et al. Value function transfer for deep multi-agent reinforcement learning based on N-step returns[C]//Proc. of the 28th International Joint Conference on Artificial Intelligence, 2019: 457–463.
25	JIN C, KRISHNAMURTHY A, SIMCHOWITZ M, et al. Reward-free exploration for reinforcement learning[C]//Proc. of the International Conference on Machine Learning, 2020: 4870–4879.
26	ZANETTE A, WAINWRIGHT M. Stabilizing Q-learning with linear architectures for provable efficient learning[C]//Proc. of the International Conference on Machine Learning, 2022: 25920–25954.
27	WANG J, ZHANG Y, KIM T K, et al. Shapley Q-value: a local reward approach to solve global reward games[C]//Proc. of the AAAI Conference on Artificial Intelligence, 2020: 7285–7292.
28	OLSEN L H B, GLAD I K, JULLUM M, et al. A comparative study of methods for estimating model-agnostic Shapley value explanations[J]. Data Mining and Knowledge Discovery, 2024, 38 (4): 1782- 1829. doi: 10.1007/s10618-024-01016-z
29	LITTMAN M L. Markov games as a framework for multi-agent reinforcement learning[C]//Proc. of the 11th International Conference on Machine Learning, 1994: 157–163.
30	VASANKARI L, SAASTAMOINEN K. Strategizing the shallows: leveraging multi-agent reinforcement learning for enhanced tactical decision-making in Littoral Naval Warfare[C]//Proc. of the International Conference on Artificial Intelligence Applications and Innovations, 2024: 129–141.
31	ZHANG K, KAKADE S, BASAR T, et al. Model-based multi-agent RL in zero-sum Markov games with near-optimal sample complexity[J]. Advances in Neural Information Processing Systems, 2020, 33 (1): 1166- 1178.
32	KURNIAWATI H. Partially observable markov decision processes and robotics[J]. Annual Review of Control, Robotics, and Autonomous Systems, 2022, 5 (1): 253- 277. doi: 10.1146/annurev-control-042920-092451
33	MOGHADDAM A R, KEBRIAEI H. Multiagent reinforcement learning for Nash equilibrium seeking in general-sum Markov games[J]. IEEE Trans. on Systems, Man, and Cybernetics: Systems, 2025, 55(1): 221–227.
34	DA-COSTA A R, PARIS D M, LUJAK M, et al. Dynamic and cooperative multi-agent task allocation: enhancing Nash equilibrium through learning[C]//Proc. of the IEEE International Conference on Systems, Man, and Cybernetics, 2024: 1860–1866.

参数	数值
智能体数量	50
作战区域/m	1000×1000
学习率α	0.1
折扣因子γ	0.9
训练轮数	1000
每次仿真任务轮数	200

算法	限制时间/s	迭代次数	平均目标函数值
本文算法	1	20	8.33
	2	50	7.118
	5	100	7.089
未改进的DQN算法	1	20	18.353
	2	50	10.108
	5	100	9.089

方法	任务完成率/%	资源利用率/%	协同效能（分数）	抗干扰鲁棒性（分数）
静态联盟策略	70.5	55	3.81	4.55
拍卖机制任务分配	65.66	60.8	5.68	5.69
强化学习协作模型	85	80.5	8.45	7.88
本文算法	98.1	92.7	10.88	9.65

[1]	尹天然, 罗贺, 石悦, 强小蝶, 王国强. 基于强化混合遗传算法的卫星与无人机联合任务规划方法[J]. 系统工程与电子技术, 2026, 48(7): 2293-2306.
[2]	杨京礼, 王莹, 高天宇, 方晓彤. 面向故障传播风险抑制的舰船电子信息系统动态重构方法[J]. 系统工程与电子技术, 2026, 48(7): 2307-2318.
[3]	李强, 周荻, 李思远, 林玉荣. 基于元学习与预测控制的导弹自适应制导[J]. 系统工程与电子技术, 2026, 48(7): 2424-2433.
[4]	马赞, 刘禹彬, 白杰, 陈勇, 孙淑光. 基于分层STPA-MC的无人机智能避让适航安全风险评估[J]. 系统工程与电子技术, 2026, 48(6): 2000-2013.
[5]	贾振帅, 肖冰, 钱寒雨, 张哲宇. 基于多策略学习的航天器在轨观测机动决策方法[J]. 系统工程与电子技术, 2026, 48(5): 1590-1598.
[6]	张华卿, 张晓飞, 郝明瑞, 姜吉祥, 李闪. 鲁棒多智能体协同对抗策略离线强化学习[J]. 系统工程与电子技术, 2026, 48(5): 1670-1681.
[7]	张少辉, 李璐璐, 李亚飞, 吴庆顺, 李冠峰, 徐明亮. 舰载机弹药保障作业规划方法研究综述[J]. 系统工程与电子技术, 2026, 48(4): 1303-1321.
[8]	金志刚, 刘泽培, 武晓栋. 物联网安全与隐私保护中博弈论的应用综述[J]. 系统工程与电子技术, 2026, 48(3): 1072-1082.
[9]	杨鹏程, 杨清清, 高盈盈, 杨志伟, 杨克巍, 艾波. 基于强化学习的海上移动目标搜索路径规划[J]. 系统工程与电子技术, 2026, 48(2): 515-523.
[10]	王旭, 蔡光斌, 余晓亚, 叶子绮, 单斌. 基于双动态PPO算法的高超声速飞行器姿态控制[J]. 系统工程与电子技术, 2026, 48(2): 694-704.
[11]	林志康, 刘甲磊, 马佳智, 施龙飞, 徐进宝. 利用分布式辐射源闪烁诱偏的抗反辐射方法[J]. 系统工程与电子技术, 2026, 48(1): 1-11.
[12]	薛锦妍, 张雅声, 陶雪峰, 杨茗棋, 赵帅龙. GEO航天器轨道机动控制研究进展[J]. 系统工程与电子技术, 2026, 48(1): 290-300.
[13]	宋传龙, 张倩武, 何健, 周文骏, 王辉, 孔巍巍, 田文波. 基于MADDPG算法的星地协同边缘计算任务卸载方法[J]. 系统工程与电子技术, 2026, 48(1): 350-360.
[14]	姚鹏, 韩美玉, 王德川, 高志诚. 基于对抗进化强化学习的多无人艇追捕方法[J]. 系统工程与电子技术, 2025, 47(9): 2960-2970.
[15]	魏潇龙, 吴亚荣, 姚登凯, 赵顾颢. 基于深度强化学习的无人机空战机动分层决策算法[J]. 系统工程与电子技术, 2025, 47(9): 2993-3003.

对抗场景下基于多智能体联盟更新的指控逻辑链优化问题研究

Research on command and control logic chain optimization based on multi-agent coalition updates in adversarial scenarios

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 4

参考文献 34

相关文章 15

编辑推荐

Metrics

本文评价