QoS-Aware Power Allocation in Mixed-Numerology NOMA

Abstract

Mixed Numerology Non-Orthogonal Multiple Access (MN-NOMA) is a promising technique for beyond-5G wireless networks, but practical deployments face two major challenges: inter-numerology interference (INI) caused by overlapping subcarriers and residual interference due to imperfect successive interference cancellation (SIC). Most existing studies address these two issues separately. This paper proposes a QoS-aware power allocation framework that jointly mitigates INI and imperfect SIC in downlink MN-NOMA systems. The resource allocation problem is formulated with a logarithmic utility function, that aims to balance the overall spectral efficiency and user fairness. The resulting optimization problem is highly non-convex because of the coupled interference terms. To solve it efficiently, an alternating optimization and successive convex approximation (AO-SCA) framework is developed, where the original problem is iteratively transformed into tractable convex sub problems. Simulation results demonstrate clear performance gains over Equal Power Allocation (EPA) and Fixed Power Allocation (FPA) schemes. The proposed framework improves spectral efficiency, particularly in the low-to-moderate SNR region, while maintaining reliable performance under practical interference conditions. Unlike schemes that favor only strong-channel users, the proposed method provides a balanced fairness-efficiency tradeoff, maintaining a Jain's fairness index of approximately 0.67 while reducing outage probability to near-zero levels at SNR values above 30 dB. These results indicate that the proposed AO-SCA framework provides an effective and practical solution for fairness-aware