Author/s: Yasser A.Al-khafaje Email: yaziz1556@gmail.com
RaedS.H.AL-Musawi Email: Raed.Almusawi@uobabylon.edu.iq
DOI: https://doi.org/10.31972/iceti2024.032
Abstract
In next-generation wireless communication, non-orthogonal multiple access (NOMA) emerges as a disruptive technology that allows several users to connect concurrently on a shared time-frequency resource using methods like successive interference cancellation (SIC). Power allocation and user clustering are the main areas of attention for this review, which looks into the optimization of NOMA systems that mostly rely on resource allocation, which is vital to improving their performance. Identifying the optimal resource distribution involves a large computational expenditure because of non-convex optimization problems. The adoption of deep learning techniques for power allocation is explored, addressing the inherent complexity of the optimization problem. Deep learning, adept at learning intricate patterns from data, is positioned as a powerful tool for overcoming computational challenges. The article emphasizes that the future evolution of NOMA-based wireless communication hinges on effectively leveraging deep learning techniques for resource allocation, outlining potential directions for future research, and highlighting deep learning’s transformative potential in addressing complex optimization problems in NOMA systems.
Keywords: Deep Learning, Resource Allocation, Power Allocation, User association, User Clustering, Non-orthogonal multiple access (NOMA), and Deep Reinforcement Learning (DRL).
Full Paper