Joint scheduling and resource allocation based on reinforcement learning in integrated access and backhaul networks

Abstract

In recent wireless networks, the integrated access and backhaul (IAB) network provides a cost-effective solution to enhance network performance and has emerged as a key technology not only for beyond 5G but also for 6G. Because of the inherent nature of IAB, access and backhaul links share the same resource pool causing cross-link interference. To address this challenge, this paper investigates an algorithm based on reinforcement learning (RL) for joint scheduling and resource allocation (RA) problem, aiming to mitigate interference and enhance user data rates. However, the scale of this joint problem is too large to solve using RL alone. Therefore, this paper proposes decomposing the joint problem into virtual scheduling and RL-based RA (RL-RA), and then solving them collaboratively. Simulation results also show that the proposed algorithm significantly improves performance and can be applied comprehensively to various duplex modes, including half and full duplex types, and different frequency bands, such as sub-6GHz and mmWave.