文献类型：期刊文献

中文题名：基于协同合作的多智能体控制系统算法探究

英文题名：Research on Multi-Agent Control System Algorithm Based on Cooperative Cooperation

作者：吴志安[1];赖永朋[2];朱有亮[1];李德荣[1]

机构：[1]广东海洋大学机械与动力工程学院,广东湛江524088;[2]广东海洋大学数学与计算机学院,广东湛江524088

年份：2022

卷号：51

期号：8

起止页码：82

中文期刊名：机电工程技术

外文期刊名：Mechanical & Electrical Engineering Technology

基金：国家级大学生创新创业训练计划项目(编号:202110566015)。

语种：中文

中文关键词：多智能体;集中式与分布式;协同控制;无人控制算法;算法分析

外文关键词：multi-agent;centralized and distributed;cooperative control;unmanned control algorithm;algorithm analysis

中文摘要：相对于单架智能机体,多智能体编队有着许多优势,如可以降低每架智能机体负载、提高任务完成效率等。而且多智能体编队兼备自主性、灵活性、稳定性,且能够高成功率、高效率地完成探测、物品运送、甚至更复杂的任务,满足不同场景的实用需求。只不过多智能体控制系统是一种复杂的动态体系,面临外部条件的不断改变和任务的灵活分配,单纯使用人工制定的方式容易造成工作效率降低,不能满足动态的变化。而且,从每个智能体的角度来看,环境是不稳定的,不利于收敛。从多智能体编队的两种思路,集中式分布与分散式分布出发,探究不同的控制算法的优缺点与适用情况。通过对比发现,分布式算法通过对每个智能机体自由度的调配,可以实现对复杂环境做出自由应变,而每个智能机体做出应答后,其他智能机体根据信息参数的改变而改变,从而整个机体编队比集中式算法有着更好的自适应性,在多智能体系统控制算法中,分布式算法远优于集中式算法。

外文摘要：Compared with single intelligent body, multi-agent formation has many advantages, such as reducing the load of each intelligent body and improving the efficiency of task completion. Moreover, multi-agent formation has autonomy, flexibility, stability, and can complete detection, goods transportation, and even more complex tasks with high success rate and high efficiency, meeting the practical needs of different scenarios. However, the multi-agent control system is a complex dynamic system, which is faced with the continuous change of external conditions and the flexible allocation of tasks. Simply using the manual method is easy to reduce the work efficiency and cannot meet the dynamic changes. Moreover, from the perspective of each agent, the environment is unstable and not conducive to convergence. Therefore, this paper explores the advantages and disadvantages and application of different control algorithms from the two ideas of multi-agent formation,centralized distribution and decentralized distribution. Through comparison, it is found that the distributed algorithm is free for each intelligent body. After each intelligent body makes a response, other intelligent bodies change according to the change of information parameters, so the whole body formation has better adaptability than the centralized algorithm. Among the multi-agent system control algorithms, the distributed algorithm is far better than the centralized algorithm.