Anytime Lifelong Multi-Agent Pathfinding in Topological Maps

This study addresses a lifelong multi-agent path finding (lifelong MAPF) problem that continuously solves an MAPF instance online according to newly assigned goals. Specifically, we focus on in topological map. is challenging because the movement of agent restricted narrow corridors map, rather than entire map area. Bypasses may be limited or farther away corridors, significantly complicating computation paths. Furthermore, low-quality solutions cause traffic congestion even deadlock corridor. Therefore, propose novel method effectively resolves conflicts based anytime strategy. gradually improves solution quality by updating sub-paths with heavy congestion. adopt several improvement steps resolve corridor conflict-based search (CBS). reduces space and time CBS. We conducted extensive experiments various maps warehouse railway environments. The results show proposed outperforms state-of-the-art methods terms throughput success rate. In particular, can collisions longer horizon existing methods, considerably improving long-range