Traffic Management for Swarm Production

In the context of swarm production, a novel pro-duction paradigm is proposed in [1]. This paper explores and benchmarks two different algorithms for multi-robot path finding as well as two different queuing strategies; one of these being the novel concept of dynamic queue positions. The algorithms are benchmarked on a physical swarm production system consisting of five mobile robots acting as carrier robots for material handling, four virtual workstations, and a central swarm manager. Testing of the traffic management strategies shows that considering the temporal dimension is better suited for swarm production, similar to solutions found for the multi-agent-path-finding and multi-agent-pickup-and-delivery problem. Further-more, dynamic queue points were found to potentially improve the scalability of a swarm production system.