Multi-Agent Pickup and Delivery in Presence of Another Team of Robots

In a Multi-Agent Pickup and Delivery (MAPD) problem, a group of agents has to accomplish subsequent pickup and delivery tasks while avoiding collisions. Tasks are provided at runtime, making MAPD a combination between classical Multi-Agent Path Finding (MAPF) and online task assignment. In this paper, we consider a new formulation of the MAPD problem, in which a guest team of agents has to solve its MAPD problem without interfering with the main team of agents, called home team, that is already carrying out its own MAPD problem in the same environment. The two teams are independent, and inter-teams communications are not allowed. We address the problem from the point of view of the guest agents, and we propose that they build a model of the behavior of the home team and exploit this information in planning their paths. Experimental results show that the inclusion of information about the behavior of home agents in guests' planning phase reduces the number of potential collisions (and hence the replanning overhead) and decreases tasks' completion time for guests.