为保证各舰船之间的协同占位或位置协调，确定各舰船需要占据的位置或区域，提出舰船多智能体协同占位方案数学建模优化方法。以舰船多智能体运动数学模型为基础，分析各个舰船智能体的运动和航行状态，判断各个智能体的航行领域，并确定舰船智能体的最近会遇距离；采用MADDPG（Multi-Agent Deep Deterministic Policy Gradient）算法结合该距离构建舰船多智能体多元组，以此获取舰船多智能体最佳的占位决策结果；在此基础上，引入同结构变换优化舰船多智能体占位编队结构，保证每个个体的精准占位以及位姿状态的一致性。测试结果表明，该方法能够有效完成各个智能体的位置决策，保证编队位姿状态的一致性，整个舰船智能体编队位置和理想位置之间的误差均低于（5, 5）m。
To ensure collaborative occupancy or position coordination among ships, determine the positions or areas that each ship needs to occupy, and propose a mathematical modeling optimization method for ship multi-agent collaborative occupancy scheme. Based on the mathematical model of ship multi-agent motion, analyze the motion and navigation status of each ship intelligent agent, determine the navigation domain of each intelligent agent, and determine the nearest encounter distance of the ship intelligent agent. Using the MADDPG (Multi Agent Deep Determining Policy gradient) algorithm combined with this distance to construct a multi-agent multi group for ships, in order to obtain the optimal occupancy decision results for ship multi-agent systems. On this basis, the same structure transformation is introduced to optimize the formation structure of ship multi-agent occupancy, ensuring accurate occupancy and consistency of pose state for each individual. The test results show that this method can effectively complete the position decision-making of various intelligent agents, ensure the consistency of the formation pose state, and the error between the position of the entire ship intelligent agent formation and the ideal position is less than (5, 5) m.
2024,46(24): 155-159 收稿日期：2024-10-12
DOI：10.3404/j.issn.1672-7649.2024.24.026
分类号：TP242
基金项目：国家自然科学基金面上项目（62372286）；上海市科委“科技创新行动计划”基础研究领域项目（23JC1403200）
作者简介：孔令彦(1989-),女,硕士,讲师,研究方向为计算数学模式识别
参考文献：
[1] 丁善婷, 王淼, 董正琼, 等. 基于多智能体的舰船装备健康状态仿真评估方法[J]. 中国机械工程, 2022, 33(10): 1169-1177.
DING S T, WANG M, DONG Z Q, et al. A multi-agent-based simulation method for health state assessments of naval equipment[J]. China Mechanical Engineering, 2022, 33(10): 1169-1177.
[2] 尹安. 基于墨子平台的水下多智能体协同占位决策算法设计[J]. 中国舰船研究, 2024, 19(S1): 109-114.
YIN A. Design and research of underwater multi-agent cooperative getting-to-the-firing-position decision-making algorithm based on MoZi platform[J]. Chinese Journal of Ship Research, 2024, 19(S1): 109-114.
[3] 范学满, 王永洁, 田立业. 基于群体智能的多UUV动态任务分配方法[J]. 舰船科学技术, 2023, 45(1): 94-100.
FAN X M, WANG Y J, TIAN L Y. A dynamic task assignment method based on swarm intelligence for multi-UUV[J]. Ship Science and Technology, 2023, 45(1): 94-100.
[4] AKDAG M, FOSSEN T I, JOHANSEN T A, et al. A decision support system for autonomous ship trajectory planning[J]. Ocean engineering, 2024, 292(1): 1.1–1.16.
[5] 王子豪, 张严心, 黄志清, 等. 部分可观测下基于RGMAAC算法的多智能体协同[J]. 控制与决策, 2023, 38(5): 1267-1277.
WANG Z H, ZHANG Y X, HUANG Z Q, et al. Multi-agent collaboration based on RGMAAC algorithm under partial observability[J]. Control and Decision, 2023, 38(5): 1267-1277.
[6] 褚晶, 李佩文, 岳颀. 基于约束优化的多智能体协同编队与避障[J]. 南京航空航天大学学报, 2024, 56(3): 545-560.
CHU J, LI P W, YUE Q. Multi-agent collaborative formation with obstacle avoidance based on constrained optimization[J]. Journal of Nanjing University of Aeronautics & Astronautics, 2024, 56(3): 545-560.
[7] 徐言民, 律建辉, 刘佳仑, 等. 基于CSSOA的多船智能避碰决策研究[J]. 中国舰船研究, 2023, 18(6): 88-96.
XU Y M, LV J H, LIU J L, et al. Multi-vessel intelligent collision avoidance decision-making based on CSSOA[J]. Chinese Journal of Ship Research, 2023, 18(6): 88-96.
[8] 黄蓉, 周军, 黄浩乾. 基于内模原理的多智能体蜂拥编队跟踪控制[J]. 计算机仿真, 2022, 39(4): 375-380.
HUANG R, ZHOU J, HUANG H Q. Multi-agent flocking formation and trajectory tracking control under internal model principle[J]. Computer Simulation, 2022, 39(4): 375-380.
