Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Delay-Oriented Distributed Scheduling with TransGNN

Boxuan Wen, Junyu Luo

TL;DR

The paper tackles delay-minimization in wireless multi-hop networks by introducing a Transformer Graph Neural Network (TransGNN) to estimate per-link utilities for scheduling, paired with a Local Greedy Solver (LGS) to ensure feasible independent sets. By replacing conventional GCN encoders with TransGNNs, the approach captures long-range interference patterns via attention and incorporates positional encoding for complex topologies. Empirical results on diverse graph models show that TransGNN–LGS reduces average queue lengths compared with LGS and GCN baselines, particularly in heterogeneous networks, with ablations confirming the critical roles of attention and encoding. The work outlines future directions including temporal dynamics, scalable attention, theoretical guarantees, and deployment in practical 5G/6G-like testbeds.

Abstract

Minimizing transmission delay in wireless multi-hop networks is a fundamental yet challenging task due to the complex coupling among interference, queue dynamics, and distributed control. Traditional scheduling algorithms, such as max-weight or queue-length-based policies, primarily aim to optimize throughput but often suffer from high latency, especially in heterogeneous or dynamically changing topologies. Recent learning-based approaches, particularly those employing Graph Neural Networks (GNNs), have shown promise in capturing spatial interference structures. However, conventional Graph Convolutional Networks (GCNs) remain limited by their local aggregation mechanism and their inability to model long-range dependencies within the conflict graph. To address these challenges, this paper proposes a delay-oriented distributed scheduling framework based on Transformer GNN. The proposed model employs an attention-based graph encoder to generate adaptive per-link utility scores that reflect both queue backlog and interference intensity. A Local Greedy Solver (LGS) then utilizes these utilities to construct a feasible independent set of links for transmission, ensuring distributed and conflict-free scheduling.

Delay-Oriented Distributed Scheduling with TransGNN

TL;DR

The paper tackles delay-minimization in wireless multi-hop networks by introducing a Transformer Graph Neural Network (TransGNN) to estimate per-link utilities for scheduling, paired with a Local Greedy Solver (LGS) to ensure feasible independent sets. By replacing conventional GCN encoders with TransGNNs, the approach captures long-range interference patterns via attention and incorporates positional encoding for complex topologies. Empirical results on diverse graph models show that TransGNN–LGS reduces average queue lengths compared with LGS and GCN baselines, particularly in heterogeneous networks, with ablations confirming the critical roles of attention and encoding. The work outlines future directions including temporal dynamics, scalable attention, theoretical guarantees, and deployment in practical 5G/6G-like testbeds.

Abstract

Minimizing transmission delay in wireless multi-hop networks is a fundamental yet challenging task due to the complex coupling among interference, queue dynamics, and distributed control. Traditional scheduling algorithms, such as max-weight or queue-length-based policies, primarily aim to optimize throughput but often suffer from high latency, especially in heterogeneous or dynamically changing topologies. Recent learning-based approaches, particularly those employing Graph Neural Networks (GNNs), have shown promise in capturing spatial interference structures. However, conventional Graph Convolutional Networks (GCNs) remain limited by their local aggregation mechanism and their inability to model long-range dependencies within the conflict graph. To address these challenges, this paper proposes a delay-oriented distributed scheduling framework based on Transformer GNN. The proposed model employs an attention-based graph encoder to generate adaptive per-link utility scores that reflect both queue backlog and interference intensity. A Local Greedy Solver (LGS) then utilizes these utilities to construct a feasible independent set of links for transmission, ensuring distributed and conflict-free scheduling.

Paper Structure

This paper contains 17 sections, 4 equations, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Theoretical Framework and Literature Study
  3. Research Questions and Hypotheses
  4. Method
  5. System Model
  6. Problem Statement
  7. Delay-Oriented Scheduling with TransGNNs
  8. Model Architecture
  9. Scheduling with Local Greedy Solver (LGS)
  10. Numerical Experiments
  11. Simulation Setup
  12. Training Protocol
  13. Implementation Details
  14. Evaluation Protocol
  15. Results and Analysis
  16. ...and 2 more sections

Figures (3)

  • Figure 1: The proposed model structure
  • Figure 2: The average queue length ratio to LGS of the proposed methods
  • Figure 3: The average queue length ratio to LGS of the GCN based methods