The characters and advantages of different algorithms in multi-robot path planning

Abstract

Multi-agent path finding (MAPF) issue has recently gained popularity, with the goal of planning for several agents who can all follow instructions at the same time without colliding. It is a common difficulty when attempting to identify pathways to several agents while ensuring that each agent reaches its destination without colliding and confusion. Real-world MAPF applications such as warehouse management, food, and beverage distribution, and aviation management are paid a number of attention from people in recent years. This article discusses centralized and decentralized techniques in MAPF challenges. Several traditional algorithms will be quoted and introduced through the help of some past papers and experiments. Finally, it was determined that decentralized strategies generally provide more powerful performance, whereas centralized strategies are generally capable of computing optimal solution sequences with minimal coupling and solving complex problems, ranging from relatively few robot problems with high coupling to many robot problems with low coupling.