Leveraging Muiti-cell NOMA for Cell Edge

Non-orthogonal multiple access (NOMA) is a promising technique for performance enhancement in cellular networks. This paper investigates how much NOMA can offer for cell edge in multi-cell scenarios. We consider the joint optimization problem of time-frequency resource allocation, user pairing, and power split, for scaling up the demand delivered for edge users. Our approach for problem solving consists in iteratively applying a fixed-point method to the cells, and, for each cell, deriving an algorithm guaranteeing optimum for single-cell optimization. By embedding the fixed-point method into bi-section search, we are able to show the overall approach guarantees global optimality. Numerical results demonstrate that multi-cell NOMA optimization has much more to offer over orthogonal multiple access for the experience of edge users, in particular for high-demand and resource-limited scenarios.