Intelligent power optimization for capacity maximization in IRS-assisted NOMA networks

This paper explores the performance of non-orthogonal multiple access (NOMA) in networks enhanced by intelligent reflecting surfaces, with a particular focus on optimizing power allocation to balance system capacity and user fairness. This article derives the optimal power allocation strategy and employs Karush–Kuhn–Tucker conditions to analytically solve the optimization problem. MATLAB simulations are used to validate the approach, ensuring both users meet their minimum data rate requirements under total power constraints. Results show that the optimal power allocation strategy prioritizes the weaker user (UE2) to ensure fairness while maintaining a sufficient capacity margin for the stronger user (UE1). As total power increases, the achievable capacity of UE1 grows logarithmically, reaching 19.41 bps/Hz at 20W, while UE2's capacity initially increases but saturates around 1.74 bps/Hz due to interference from UE1. Consequently, total system capacity exhibits diminishing returns, increasing from 19.122 bps/Hz at 5W to 21.132 bps/Hz at 20W. Further analysis reveals that as UE1's power increases, interference to UE2 rises, limiting its capacity growth. Despite this, optimization ensures both users meet their minimum capacity requirements, demonstrating the efficiency of NOMA in managing power distribution and interference.