Joint Power Allocation and Antenna Selection for Network Sum-Rate Maximization in Clustered Downlink NOMA Networks

This paper considers the problem of joint power allocation and antenna selection (J-PA-AS) for downlink clustered non-orthogonal multiple-access (NOMA) networks. In particular, the goal is to perform antenna selection for each user cluster and allocate transmit power to its users so as to maximize the network sum-rate, while satisfying quality-of-service (QoS) requirements. The formulated problem happens to be non-convex and NP-hard, and thus, there is no computationally-efficient approach to solve it directly. In turn, a low-complexity two-stage algorithm is proposed, where the first stage optimally solves the sum-rate maximizing power allocation problem for each (antenna, user cluster) pair. In the second stage, antenna selection is optimally solved in polynomial-time complexity via the Kuhn-Munkres with backtracking (KMB) algorithm. Simulation results are provided to validate the proposed algorithm, which is shown to efficiently yield the optimal network sum-rate, and in comparison to the optimal J-PA-AS scheme (solved via a global optimization package), and superior to other benchmark schemes. The impact of spatial-diversity on the network sum-rate is also highlighted, where it is shown that the greater the number of antennas at the base-station is, the higher the network sum-rate, and the lower the outage events.