Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

GNN-Based Joint Channel and Power Allocation in Heterogeneous Wireless Networks

Lili Chen, Jingge Zhu, Jamie Evans

TL;DR

The paper tackles the non-convex problem of jointly allocating channels and transmit power in heterogeneous wireless networks. It introduces JCPGNN, a heterogeneous-graph neural network with a shared message-passing layer and two task-specific heads to predict channel assignments and power levels simultaneously, enabling channel reuse and interference-aware optimization. Empirical results show JCPGNN achieves near-optimal throughput (about 95% of the exhaustive-optimal) on moderate-scale scenarios, generalizes well to larger networks, and remains robust when CSI is corrupted, all with significantly lower computational cost than traditional methods. This approach demonstrates potential for real-time, scalable resource management in 5G-and-beyond environments, leveraging topology-aware learning without requiring labeled optimal solutions.

Abstract

The optimal allocation of channels and power resources plays a crucial role in ensuring minimal interference, maximal data rates, and efficient energy utilisation. As a successful approach for tackling resource management problems in wireless networks, Graph Neural Networks (GNNs) have attracted a lot of attention. This article proposes a GNN-based algorithm to address the joint resource allocation problem in heterogeneous wireless networks. Concretely, we model the heterogeneous wireless network as a heterogeneous graph and then propose a graph neural network structure intending to allocate the available channels and transmit power to maximise the network throughput. Our proposed joint channel and power allocation graph neural network (JCPGNN) comprises a shared message computation layer and two task-specific layers, with a dedicated focus on channel and power allocation tasks, respectively. Comprehensive experiments demonstrate that the proposed algorithm achieves satisfactory performance but with higher computational efficiency compared to traditional optimisation algorithms.

GNN-Based Joint Channel and Power Allocation in Heterogeneous Wireless Networks

TL;DR

The paper tackles the non-convex problem of jointly allocating channels and transmit power in heterogeneous wireless networks. It introduces JCPGNN, a heterogeneous-graph neural network with a shared message-passing layer and two task-specific heads to predict channel assignments and power levels simultaneously, enabling channel reuse and interference-aware optimization. Empirical results show JCPGNN achieves near-optimal throughput (about 95% of the exhaustive-optimal) on moderate-scale scenarios, generalizes well to larger networks, and remains robust when CSI is corrupted, all with significantly lower computational cost than traditional methods. This approach demonstrates potential for real-time, scalable resource management in 5G-and-beyond environments, leveraging topology-aware learning without requiring labeled optimal solutions.

Abstract

The optimal allocation of channels and power resources plays a crucial role in ensuring minimal interference, maximal data rates, and efficient energy utilisation. As a successful approach for tackling resource management problems in wireless networks, Graph Neural Networks (GNNs) have attracted a lot of attention. This article proposes a GNN-based algorithm to address the joint resource allocation problem in heterogeneous wireless networks. Concretely, we model the heterogeneous wireless network as a heterogeneous graph and then propose a graph neural network structure intending to allocate the available channels and transmit power to maximise the network throughput. Our proposed joint channel and power allocation graph neural network (JCPGNN) comprises a shared message computation layer and two task-specific layers, with a dedicated focus on channel and power allocation tasks, respectively. Comprehensive experiments demonstrate that the proposed algorithm achieves satisfactory performance but with higher computational efficiency compared to traditional optimisation algorithms.
Paper Structure (15 sections, 8 equations, 6 figures, 1 table, 1 algorithm)

This paper contains 15 sections, 8 equations, 6 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Resource Allocation problems
  3. System Model
  4. graph neural networks for JCPA problems
  5. Graph Modelling in Wireless Networks
  6. JCPA on Heterogeneous Graph Neural Networks
  7. Message Computation Layer
  8. Task-specific Layer
  9. Experiments and Results
  10. Simulations Setup
  11. Performance Comparison
  12. Generalisation Capability
  13. Robustness to Corrupted CSI
  14. Time Complexity
  15. conclusion

Figures (6)

  • Figure 1: Heterogeneous wireless network with different channels.
  • Figure 2: Heterogeneous graph representation of the system model.
  • Figure 3: The structure of JCPGNN.
  • Figure 4: The average sum rate for two channels scenario.
  • Figure 5: Normalised performance of JCPGNN with difference percentage of corrupted CSI input.
  • ...and 1 more figures