Distributed Link Sparsification for Scalable Scheduling Using Graph Neural Networks (Journal Version)
This addresses congestion and energy issues in wireless networks, but it is incremental as it builds on existing scheduling protocols with a new optimization method.
The paper tackles the problem of signaling overhead in dense wireless networks by proposing a distributed link sparsification scheme using graph neural networks to reduce scheduling overhead for delay-tolerant traffic while maintaining network capacity, achieving reduced congestion and radio footprints in simulations with up to 500 links.
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.