未知环境中多Agent自主协作规划策略

针对多智能体(Agent)系统在未知环境中自主协作规划存在任务死锁及协作效率不高的问题，提出一种基于改进蚁群算法的多Agent协作策略,并用于多Agent协作搬运中。该方法将Agent所处位置和目标任务之间的距离以及信息素控制因子引入蚁群算法。实验结果表明,该方法相比没有引入距离因子的协作方法，协作效率更高；相比没有引入控制因子的协作方法，可有效防止任务死锁发生。

Multi-Agent autonomous cooperation planning strategy in unknown environment

A multi-Agent cooperation strategy based on the improved ant algorithm is proposed for solving the problems of task deadlock and low cooperation efficiency when the multi-Agent system works in unknown environment, then the method is applied to multi-Agent cooperative transportation. The distance between the Agent’s position and the task’s position and the pheromone control factor are introduced in the ant colony algorithm. Experimental results show that the cooperation efficiency of this method is higher than those ones without considering the distance factor, and can avoid task deadlock more effectively than those ones without the task pheromone control factor.