Pairwise is Not Enough: Hypergraph Neural Networks for Multi-Agent Pathfinding

TL;DR

This paper introduces HMAGAT, a hypergraph neural network that captures group interactions in multi-agent pathfinding, outperforming existing pairwise methods and emphasizing the importance of higher-order representations.

Contribution

We propose HMAGAT, a hypergraph-based neural network architecture that explicitly models group dynamics in multi-agent pathfinding, achieving state-of-the-art results with fewer parameters and less training data.

Findings

HMAGAT outperforms current state-of-the-art MAPF solvers.

Hypergraph representations reduce attention dilution in dense environments.

Group interactions are crucial for effective multi-agent coordination.

Abstract

Multi-Agent Path Finding (MAPF) is a representative multi-agent coordination problem, where multiple agents are required to navigate to their respective goals without collisions. Solving MAPF optimally is known to be NP-hard, leading to the adoption of learning-based approaches to alleviate the online computational burden. Prevailing approaches, such as Graph Neural Networks (GNNs), are typically constrained to pairwise message passing between agents. However, this limitation leads to suboptimal behaviours and critical issues, such as attention dilution, particularly in dense environments where group (i.e. beyond just two agents) coordination is most critical. Despite the importance of such higher-order interactions, existing approaches have not been able to fully explore them. To address this representational bottleneck, we introduce HMAGAT (Hypergraph Multi-Agent Attention Network), a novel architecture that leverages attentional mechanisms over directed hypergraphs to explicitly capture group dynamics. Empirically, HMAGAT establishes a new state-of-the-art among learning-based MAPF solvers: e.g., despite having just 1M parameters and being trained on 100$\times$ less data, it outperforms the current SoTA 85M parameter model. Through detailed analysis of HMAGAT's attention values, we demonstrate how hypergraph representations mitigate the attention dilution inherent in GNNs and capture complex interactions where pairwise methods fail. Our results illustrate that appropriate inductive biases are often more critical than the training data size or sheer parameter count for multi-agent problems.