基于墨子平台的水下多智能体协同占位决策算法设计  

Design and research of underwater multi-agent cooperative getting-to-the-firing-position decision-making algorithm based on MoZi platform

作　　者：尹安[1] YIN An(China Ship Development and Design Center,Wuhan 430064,China)

机构地区：[1]中国舰船研究设计中心,湖北武汉430064

出　　处：《中国舰船研究》2024年第S01期109-114,共6页Chinese Journal of Ship Research

基　　金：国家部委基金资助项目。

摘　　要：[目的]开展基于墨子平台的水下战场多智能体协同占位决策算法设计研究,探索水下无人平台作战应用。[方法]提出基于墨子平台的设计思路。针对算法处理确定性规则下的行动决策,提出基于知识图谱的设计方法。采用Neo4J图数据库构建全局知识库,并设计数据交换接口,为智能算法模块提供确定性规则的辅助决策信息输入;针对算法处理非确定性和不完全信息下的行动决策,在设定场景下论述基于MADDPG的协同占位决策算法设计。提出状态空间、动作空间、奖励函数以及与知识图谱模块的交互设计思路。[结果]基于上述思路,完成了算法框架总体设计。在墨子平台中实现了知识图谱和智能算法的整合,并通过墨子平台提供的可视化手段,进一步验证了确定性规则下的行为约束与处理非确定性态势信息的智能算法进行整合的有效性。[结论]仿真结果表明,本文研究工作可为后续进一步开展多智能体协同占位算法设计提供参考。[Objectives]This study focuses on the multi-agent cooperative getting-to-the-firing-position decision-making algorithm design of underwater battlefields based on the MoZi platform,and explores the combat application of underwater unmanned platforms.[Methods]A design idea based on the MoZi platform is proposed.For action decision-making under deterministic rules,the design of knowledge graph modules is discussed.The Neo4J database is used to build a global knowledge database,and a data exchange interface is designed to provide the auxiliary decision information input of deterministic rules for the intelligent algorithm module.For action decision-making under uncertain and incomplete information,the design of a cooperative space occupying a decision-making algorithm based on MADDPG is discussed under the set scenario,and the design concept of a state space,action space,reward function and interaction with the knowledge graph module is put forward.[Results]The overall design of the algorithm framework is completed on the basis of the abovementioned ideas.The integration of the knowledge graph and intelligent algorithm is realized on the MoZi platform,and the effectiveness of the integration of behavior constraints under deterministic rules and the intelligent algorithm for processing uncertain situation information is further verified through the visualization provided by the platform.[Conclusions]This paper can provide valuable references for the design of multi-agent underwater getting-to-the-firing-position cooperative algorithms.

关 键 词：多智能体 墨子平台 知识图谱 MADDPG 

分 类 号：U662.9[交通运输工程—船舶及航道工程]