为解决海上编队协同防空作战中多任务多平台的协同决策问题，提出基于遗传模糊逻辑树的协同防空作战规则反演方法。设计面向威胁判断、目标分配、火力控制等指控决策的级联式模糊推理系统，采用演化学习从博弈对抗中学习和反演协同作战规则。作战仿真测试表明，该方法能够适应战场的动态变化且决策时延低于1 s。基于遗传模糊逻辑树的作战规则反演缓解了深度强化学习等面临的可解释性问题和作战仿真中的奖励稀疏问题，同时反演生成的作战规则也为剖析战场规律提供了新的知识来源。
To solve the multi-task and multi-platform cooperative decision-making problem in the maritime formation cooperative air defense operation, a method of inversion of cooperative air defense operation rules based on genetic fuzzy logic tree was proposed. A cascading fuzzy inference system is designed for threat assessment, target weapon assignment, fire control and other C2 tasks. Evolutionary learning is used to learn and invert cooperative combat rules from Wargaming. Based on the simulation results, the method can adapt to the dynamic changes of the battlefield and the decision-making delay is less than 1 second. The inversion of combat rules based on genetic fuzzy logic tree alleviates the lack of interpretability faced by deep reinforcement learning and the problem of reward sparsity in simulation environment. At the same time, the combat rules generated by inversion also provide a new source of knowledge for analyzing the nature of naval warfare.
2024,46(8): 180-184 收稿日期：2023-5-11
DOI：10.3404/j.issn.1672-7649.2024.08.034
分类号：E11
作者简介：李洋(1989-),男,博士,工程师,研究方向为军事运筹与仿真
参考文献：
[1] 吴勤. 美军分布式作战概念发展分析[J/OL]. 军事文摘, 2016(13): 44-47.
[2] 周玺. 未来海空分布式作战构想与力量运用初探[J/OL]. 中国电子科学研究院学报, 2020, 15(9): 856-860.
[3] 王肖飞, 严建钢, 丁伟锋, 等. 舰艇编队协同防空作战体系研究[J]. 舰船电子工程, 2011, 31(7): 1-4+62.
[4] 李烨, 邱志明, 郭勇. 编队多平台协同防空作战系统逻辑结构分析[J]. 指挥控制与仿真, 2013, 35(6): 12-16.
[5] 万里鹏, 兰旭光, 张翰博, 等. 深度强化学习理论及其应用综述[J]. 模式识别与人工智能, 2019, 32(1): 67-81.
[6] 孙彧, 曹雷, 陈希亮, 等. 多智能体深度强化学习研究综述[J]. 计算机工程与应用, 2020, 56(5): 13-24.
[7] 朱建文, 赵长见, 李小平, 等. 基于强化学习的集群多目标分配与智能决策方法[J]. 兵工学报, 2021, 42(9): 2040-2048.
[8] 王小光, 胡荣, 梁文洋. 无人机群协同作战目标分配研究综述[J]. 军事文摘, 2021(7): 32-35.
[9] 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/OL]. 软件学报, 2021: 1-17.
[10] 罗荣, 王亮, 肖玉杰, 等. 深度学习技术在军事领域应用[J]. 指挥控制与仿真, 2020, 42(1): 1-5.
[11] CORDÓN O. Genetic fuzzy systems: evolutionary tuning and learning of fuzzy knowledge bases[M]. Singapore: World Scientific, 2001.
[12] ERNEST N, CARROLL D, SCHUMACHER C, et al. Genetic fuzzy based artificial intelligence for unmanned combat aerial vehicle control in simulated air combat missions[J]. Journal of Defense Management, 2016, 6(1): 2167-0374.1000144.
[13] CABRERA I P, CORDERO P, OJEDA-ACIEGO M, et al. Fuzzy logic，soft computing，and applications [C/OL]//CABESTANY. Bio-Inspired Systems: Computational and Ambient Intelligence. Berlin, Heidelberg: Springer, 2009: 236-244.
[14] SHI Y, MIZUMOTO M, YUBAZAKI N, et al. A learning algorithm for tuning fuzzy rules based on the gradient descent method[C/OL]//Proceedings of IEEE 5th International Fuzzy Systems. New Orleans, LA, USA: IEEE, 1996, (1): 55-61.
[15] BEDE B. Mathematics of fuzzy sets and fuzzy logic[M]: Berlin, Heidelberg: Springer Berlin Heidelberg, 2013, 295: 978.3.642
[16] HERRERA F. Genetic fuzzy systems: taxonomy, current research trends and prospects[J/OL]. Evolutionary Intelligence, 2008, 1(1): 27-46.
[17] ERNEST N D. Genetic fuzzy trees for intelligent control of unmanned combat aerial vehicles[D]. University of Cincinnati, 2015.
[18] SRINIVAS M, PATNAIK L M. Adaptive probabilities of crossover and mutation in genetic algorithms[J/OL]. IEEE Transactions on Systems, Man, and Cybernetics, 1994, 24(4): 656-667.
[19] 刘朝阳, 穆朝絮, 孙长银. 深度强化学习算法与应用研究现状综述[J]. 智能科学与技术学报, 2020, 2(4): 314-326.
