Distributed Link Sparsification for Scalable Scheduling Using Graph Neural Networks (Journal Version)

2025-09-05

AI Summary

To address the high signaling overhead induced by distributed link scheduling in dense wireless networks, this paper proposes a graph neural network (GNN)-based distributed link sparsification method. The approach jointly models network topology and traffic characteristics, and employs a constraint-aware unsupervised learning framework to dynamically optimize per-link contention thresholds offline—thereby enabling low-success-probability links to voluntarily withdraw from scheduling competition while preserving network capacity. Its key innovation lies in the first integration of GNNs with unsupervised learning for link sparsification decisions, ensuring compatibility with multiple mainstream distributed scheduling protocols. In simulations of a 500-link multihop network, the method effectively alleviates congestion, reduces coverage disparity, and achieves an average 32.7% reduction in signaling overhead across four representative protocols.

Technology Category

• Planning, Routing, and Scheduling: Scheduling under Uncertainty
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization
• Machine Learning: Graph-based Machine Learning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

In wireless networks characterized by dense connectivity, the significant signaling overhead generated by distributed link scheduling algorithms can exacerbate issues like congestion, energy consumption, and radio footprint expansion. To mitigate these challenges, we propose a distributed link sparsification scheme employing graph neural networks (GNNs) to reduce scheduling overhead for delay-tolerant traffic while maintaining network capacity. A GNN module is trained to adjust contention thresholds for individual links based on traffic statistics and network topology, enabling links to withdraw from scheduling contention when they are unlikely to succeed. Our approach is facilitated by a novel offline constrained {unsupervised} learning algorithm capable of balancing two competing objectives: minimizing scheduling overhead while ensuring that total utility meets the required level. In simulated wireless multi-hop networks with up to 500 links, our link sparsification technique effectively alleviates network congestion and reduces radio footprints across four distinct distributed link scheduling protocols.

Problem

• Reducing signaling overhead in dense wireless networks
• Minimizing scheduling overhead while maintaining network utility
• Alleviating congestion and reducing radio footprints via link sparsification

Innovation

• Uses GNNs for distributed link sparsification
• Employs unsupervised learning to balance objectives
• Adjusts contention thresholds based on network topology