Performance Analysis and SINR-Based Power Allocation Strategies for Downlink NOMA Networks

This paper analyzes the performance of downlink non-orthogonal multiple-access (NOMA) networks over independent but not necessarily identically distributed Rayleigh fading channels by deriving closed-form expressions for the average signal-tointerference-plus-noise ratio (SINR), average achievable rate, and outage probability of all network users. Moreover, SINR-based power allocation strategies are studied. Specifically, the optimization problems of max-min SINR, multi-objective SINR, sum-SINR and proportional-fairness-SINR maximization under minimum SINR constraints are formulated and shown to be non-convex. Additionally, the problems of total power and sum-SINR minimization are also formulated. Particularly, the formulated problems take the form of linear-fractional programming problems. By applying intelligent reformulation techniques, these problems have been reformulated into convex optimization problems, and thus efficiently solved. Numerical evaluations are presented to validate the derived closed-form expressions for the different performance metrics, which are found to be in agreement with the network simulation results. More importantly, it has been demonstrated that the user with the best channel conditions achieves full diversity order, whereas all the other users achieve diversity orders equivalent to their ordered channel conditions. Finally, the reformulated power allocation strategies are evaluated, and shown to coincide with the original problem formulations.