Joint User Association and Bandwidth Assignment for Digital Twin-Assisted Multi-RAT Networks

TL;DR

This paper introduces a digital twin-assisted approach for joint user association and bandwidth allocation in multi-RAT networks, significantly improving sum-rates and QoS with reduced overhead and high computational efficiency.

Contribution

It proposes a novel digital twin-based heuristic algorithm for joint user association and bandwidth assignment, reducing CSI overhead in multi-RAT networks.

Findings

Up to 43% increase in sum-rate over baseline methods.

Less than 5% deviation from the theoretical optimal solution.

Achieves up to 43% improvement in QoS.

Abstract

In this paper, we investigate user equipment (UE)-radio access technology (RAT) association and bandwidth assignment to maximize sum-rates in a multi-RAT network. To this end, we formulate an optimization problem that jointly addresses UE association and bandwidth allocation, adhering to practical constraints. Because of the NP-hard nature of this problem, finding a globally optimal solution is computationally infeasible. To address this challenge, we propose a centralized and computationally efficient heuristic algorithm that aims to maximize sum-rates while enhancing quality of service (QoS). Yet, the proposed approach requires global channel state information (CSI) for near-optimal performance, which incurs substantial overhead and data collection costs in large-scale multi-RAT networks. To alleviate this burden, we use a digital twin (DT) of the multi-RAT network, leveraging its context-awareness to acquire global CSI with reduced overhead. Our numerical results reveal that our approach improves sum-rates by up to 43% over baseline method, with less than a 5% deviation from the theoretical optimal solution, while achieving up to a 43% improvement in QoS. Further analysis reveals that our method not only surpasses the optimal solution in terms of QoS enhancement, but also ensures significant computational efficiency.