基于协作协进化的多智能体机器人协作研究

协作问题一直是自主多智能体机器人系统研究的关键问题之一。基于多智能体机器人系统的CCP协作协议所生成的各智能体机器人的任务序列依赖于目标的初始顺序,因此难以得到最优解。文章提出了利用协作协进化来实现多智能体机器人之间协作的一种机制。该方法采用基于协作种群的技术来生成多智能体机器人任务执行序列,在给定的任务分解产生的所有可能解中寻找最优解,并通过交换局部知识和并行决策等手段来优化系统的性能。利用该机制,对3个智能体协作搬运8个物体进行计算机模拟,结果表明,该机制在优化任务执行序列方面作用明显,从而能有效提高多智能体机器人系统的性能。

A Study of Multi-Agent Cooperation Mechanism Based on Cooperative Co-evolution

Cooperation is one of the key problems of the automatic multi-agent robotic system all along.Since the task sequence created by CCP of the multi-agent robotic system lay on the initiated sequence of the objects,it is difficult to create the optimized solutions.This paper proposes a mechanism which realizes the cooperation of multi-agent by coop-erative co-evolution.The technique based on cooperative populations is used by the method to create the implementing sequence of the multi-agent task,and the performance of the system is optimized by exchanging local information and making decision in parallel.A experiment that three agents move eight objects by cooperation has been done.Its result shows that the mechanism can optimize the implementing sequence of the task evidently,and improve the performance of the system effectively.