Advancing Learnable Multi-Agent Pathfinding Solvers with Active Fine-Tuning
This work addresses scalable pathfinding for multi-robot systems in applications like logistics and search-and-rescue, representing an incremental advance over prior learning-based solvers.
The paper tackles the problem of multi-agent pathfinding (MAPF) by introducing MAPF-GPT-DDG, a novel approach that fine-tunes a pre-trained model with centralized expert data using a delta-data generation mechanism, resulting in improved solution quality and scalability up to 1 million agents.
Multi-agent pathfinding (MAPF) is a common abstraction of multi-robot trajectory planning problems, where multiple homogeneous robots simultaneously move in the shared environment. While solving MAPF optimally has been proven to be NP-hard, scalable, and efficient, solvers are vital for real-world applications like logistics, search-and-rescue, etc. To this end, decentralized suboptimal MAPF solvers that leverage machine learning have come on stage. Building on the success of the recently introduced MAPF-GPT, a pure imitation learning solver, we introduce MAPF-GPT-DDG. This novel approach effectively fine-tunes the pre-trained MAPF model using centralized expert data. Leveraging a novel delta-data generation mechanism, MAPF-GPT-DDG accelerates training while significantly improving performance at test time. Our experiments demonstrate that MAPF-GPT-DDG surpasses all existing learning-based MAPF solvers, including the original MAPF-GPT, regarding solution quality across many testing scenarios. Remarkably, it can work with MAPF instances involving up to 1 million agents in a single environment, setting a new milestone for scalability in MAPF domains.