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

2026-02-06

AI Summary

This work addresses the limitations of conventional graph neural networks in multi-agent path finding (MAPF), which struggle to model high-order coordination in dense scenarios due to their reliance on pairwise interactions, often yielding suboptimal solutions and diluted attention. To overcome this, we propose HMAGAT—the first directed hypergraph-based multi-agent attention network—that explicitly captures high-order interactions, thereby transcending the constraints of pairwise message passing. Remarkably, HMAGAT outperforms the current state-of-the-art learning-based MAPF solver, which uses 85 million parameters, despite employing only 1 million parameters and 1% of the training data. Attention analysis further validates the efficacy of high-order relational modeling, demonstrating that well-designed inductive biases are more critical than model scale or data volume.

Technology Category

• Multiagent Systems: Multiagent Planning
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

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.

Problem

• Multi-Agent Path Finding
• Hypergraph Neural Networks
• Higher-order Interactions
• Attention Dilution
• Group Coordination

Innovation

• Hypergraph Neural Networks
• Multi-Agent Path Finding
• Attention Mechanism
• Higher-order Interactions
• Inductive Bias