MA2CL: Multi-Agent Actor-Critic Learning Scheme for Efficient Resource Management in 5G-Enabled NB-IoT Networks

Abstract

The allocation of spectrum resources for the future 5G-enabled Narrowband Internet of Things (NB-IoT) is one of the most critical issues that need to be resolved. Due to the massive amount of data that will be generated by the IoT, the need for efficient allocation of resources is also immense. This paper presents a novel interference model for managing the allocation of resources and avoiding overlapping interference in the 5G-enabled NB-IoT Networks. It adopts Reinforcement Learning (RL)-based algorithms to improve the network throughput and prevent overlapping interference. The proposed method utilizes a Multi-Agent Actor-Critic Learning (MA²CL) algorithm, which can improve the efficiency of the network. The simulation result illustrates the prominent enhancement in the throughput and spectral efficiency of the network. The performances of the proposed algorithm have been compared with benchmark schemes and achieved a 33.3% increase in network throughput and a 26.67% boost in spectral efficiency, respectively. The proposed work for efficient NB-IoT resource management may be suitable in industrial automation and intelligent transportation systems.