Graph Attention-Guided Search for Dense Multi-Agent Pathfinding

2025-10-20

AI Summary

Real-time computation of near-optimal solutions remains challenging in dense multi-agent pathfinding (MAPF). Method: This paper proposes a tightly integrated learning-and-search hybrid framework. It innovatively embeds a multi-agent graph attention network (MAGAT) into the LaCAM search framework to construct a neural-augmented heuristic model; introduces a pretraining-fine-tuning paradigm to enhance generalization; and incorporates a lightweight deadlock detection mechanism to mitigate uncertainty arising from neural guidance. Contribution/Results: Experiments demonstrate that the framework significantly outperforms both pure search and pure learning baselines in highly coupled, dense scenarios, achieving a superior trade-off between solution quality and runtime efficiency. The results validate the effectiveness and robustness of learning-guided search for complex cooperative pathfinding tasks.

Technology Category

• Search and Optimization: Learning to Search
• Multiagent Systems: Multiagent Planning
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Finding near-optimal solutions for dense multi-agent pathfinding (MAPF) problems in real-time remains challenging even for state-of-the-art planners. To this end, we develop a hybrid framework that integrates a learned heuristic derived from MAGAT, a neural MAPF policy with a graph attention scheme, into a leading search-based algorithm, LaCAM. While prior work has explored learning-guided search in MAPF, such methods have historically underperformed. In contrast, our approach, termed LaGAT, outperforms both purely search-based and purely learning-based methods in dense scenarios. This is achieved through an enhanced MAGAT architecture, a pre-train-then-fine-tune strategy on maps of interest, and a deadlock detection scheme to account for imperfect neural guidance. Our results demonstrate that, when carefully designed, hybrid search offers a powerful solution for tightly coupled, challenging multi-agent coordination problems.

Problem

• Developing hybrid framework for dense multi-agent pathfinding optimization
• Integrating learned neural heuristics with search-based algorithms
• Solving tightly coupled multi-agent coordination in dense scenarios

Innovation

• Integrates graph attention neural policy into search algorithm
• Uses pre-train-then-fine-tune strategy on target maps
• Implements deadlock detection for imperfect neural guidance