Deciding who does what
Beyond moment-to-moment collision avoidance, multi-robot systems often face a higher-level problem: given a set of tasks, like several delivery stops or several warehouse shelves that need restocking, and a set of available robots, which robot should do which task. This is the multi-robot task allocation problem, and the choices made here directly affect how efficiently the fleet operates, whether measured by total completion time, total distance traveled, or how evenly the workload is spread across the fleet.
A naive approach, like always assigning the next task to the nearest idle robot, is easy to compute but can produce poor global outcomes. It might repeatedly load one robot near a busy area while another sits idle far away, or create long delivery times for tasks that get consistently skipped over because closer robots keep grabbing the nearby tasks first. Good allocation has to weigh distance, current workload, and fairness together, not just optimize the next single assignment in isolation.