Abstract

In this paper, we investigate the impact of user pairing on the power consumption of 2-user non-orthogonal multiple access (NOMA) systems in the downlink. We formulate the joint power allocation and user pairing problem as a mixed-integer programming problem with the objective of minimizing the total transmit power consumption. While the pairwise power allocation strategy is straightforward, for a system with $2K$ users and K NOMA pairs, there exist ${}\frac {(2K)!}{2^{K} \times K!}$ possible pairing strategies, resulting in a combinatorial search space that grows drastically with the number of users in the system. Hence, we propose an analytical approach to obtain the globally optimum user pairing strategy. Notably, our procedure has a linear time complexity of $\mathcal {O}(2K)$ , which is a significant improvement over the suboptimal and computationally expensive methods in the existing literature. We demonstrate through extensive simulations that the proposed optimal pairing strategy can attain considerable performance gains in terms of power savings compared to benchmark schemes. In particular, in a typical deployment environment, 63% of the total power budget is saved at a mean received signal-to-noise ratio (SNR) of 15.7 dB among the users. Finally, we evaluate the energy efficiency (EE) of NOMA transmission compared to the EE achieved through orthogonal multiple access (OMA) transmission. We demonstrate that the EE gain of NOMA transmission compared to OMA is improved more than sixfold at convergence by adopting the power minimization approach studied in this work, rather than adopting the sum rate maximization approach found in the literature.