Optimal User Pairing for Downlink Non-Orthogonal Multiple Access (NOMA)

In this letter, we explore user pairing in a downlink non-orthogonal multiple access (NOMA) network. As power allocation inherently intertwines with user pairing, a joint user pairing and power allocation problem is considered to optimize the achievable sum rate (ASR) with minimum rate constraint for each user, which is a mixed integer programming problem. To solve this non-convex problem, we first obtain the optimal power allocation in an NOMA system with only 2 users; then analyze the user pairing problem in a simplified situation, i.e., an NOMA system with four users. Finally, we obtain the closed-form globally optimal solution in a general NOMA system. Extensive performance evaluations are conducted to compare the ASRs of the NOMA and OMA systems. Results show that the performance of the NOMA system with the proposed optimal user pairing is significantly better than that of the OMA system, as well as the performance of the NOMA system with random user pairing.