From Gridworlds to Warehouses: Adapting Lightweight One-shot Multi-Agent Pathfinding for AGVs
For warehouse automation researchers, this provides a more realistic MAPF formulation and baseline algorithm comparisons, though the adaptation is incremental.
This work introduces multi-agent warehouse pathfinding (MAWPF), a practical MAPF variant for differential-drive AGVs with constraints like multi-step rotations and acceleration, and adapts suboptimal algorithms (PP, LNS2, PIBT, LaCAM). Benchmarking shows PIBT-based methods scale well but with higher costs, while PP and LNS2 struggle with many agents.
Multi-agent pathfinding (MAPF) under one-shot planning is a core component of warehouse automation, yet classical formulations typically assume four-connected 2D grids with unit-time moves in four directions. To fill reality gaps while still being trackable with discrete combinatorial search, this work proposes a more practical counterpart tailored to differential-drive AGVs. We term this multi-agent warehouse pathfinding (MAWPF), featured with four constraints: (i) agent actions are restricted to straight motion and in-place rotation; (ii) rotations require multi-step costs; (iii) acceleration and deceleration are considered, and; (iv) follower collisions are prohibited to prevent rear-end crashes. To solve MAWPF efficiently, we adapt representative suboptimal MAPF algorithms-PP, LNS2, PIBT, and LaCAM-and conduct comprehensive benchmarking. Our experiments reveal that PP and LNS2 struggle to solve instances with many agents, while PIBT-based approaches achieve preferable scalability with increased solution cost. We believe that these constitute an important step toward adapting classical gridworld MAPF to operational warehouse setups.