Unsupervised Graph-based Learning Method for Sub-band Allocation in 6G Subnetworks

In this paper, we present an unsupervised approach for frequency sub-band allocation in wireless networks using a graph-based learning method. We consider a scenario of dense deployment of subnetworks in the factory environment. The limited number of sub-bands must be optimally allocated to coordinate inter-subnetwork interference. Traditional iterative solutions may not scale to the large scale and density of subnetwork deployment due to their execution overhead limitations. Hence, we consider a data-driven approach based on graph neural networks.
We model the subnetwork deployment as an interference graph and propose an unsupervised learning approach to optimize the sub-band allocation using graph neural networks. This approach is inspired by the graph colouring heuristic and the Potts model. The numerical evaluation shows that the proposed method achieves close performance to the centralized greedy colouring sub-band allocation heuristic with lower computational time complexity. In addition, it incurs reduced signalling overhead compared to iterative optimization heuristics that require all the mutual interfering channel information. We further demonstrate that the method is robust to different network settings.