AI-Based Resource Allocation for Energy and Spectrum Efficiency in 6G Networks

Abstract

Another paradigm shift that is in the process of being introduced in the telecommunication sector is the sixth generation (6G) wireless networks whereby a low degree of latency of operation, high spectral performance and energy-conscious design are being introduced to meet the constantly increasing number of devices and data requirements of the air. Although it has been revealed that there are different papers on the utilization of AI in 6G, the majority of them remark on two phenomena that cannot be correlated with one another and these include spectrum and energy efficiency. The systematic and mutual execution of AI-based resource allocation will be used to address the gap in the paper through the joint maximization of the energy efficiency and spectral efficiency of 6G networks.The paper proposes a new taxonomy of AI techniques that is determined by the deployment layer (edge, core, and THz frontier), the most popular known architectural and algorithmic operators of the joint optimization, and a comparative analysis of the known available methods of the technological fields, e.g., adaptive modulation, handover optimization, meta-learning, federated learning and explainable AI in comparison with the available surveys. It further possesses an offer of synthesis of digital twins on a resource-centric approach, blockchain-based Internet of Things, and secure communication protocols.Key contributions include: (1) a unified classification of AI approaches targeting co-optimization of spectrum and energy; (2) a comparative evaluation of techniques across key 6G scenarios (UDN, IoT, THz, RIS); and (3) identification of open research challenges and practical deployment constraints. The paper further outlines how AI fosters sustainability, scalability, and transparency in next-generation wireless systems.