Power Minimization Resource Allocation for Underlay MISO-NOMA SWIPT System

The combination of cognitive radio and non-orthogonal multiple access (NOMA) has tremendous potential to achieve high spectral efficiency in the IoT era. In this paper, we focus on the energy-efficient resource allocation of a cognitive multiple-input single-output NOMA system with the aid of simultaneous wireless information and power transfer. Specifically, a non-linear energy harvesting (EH) model is adopted to characterize the non-linear energy conversion property. In order to achieve the green design goal, we aim for the minimization of the system power consumption by jointly designing the transmit beamformer and the receive power splitter subject to the information transmission and EH harvesting requirements of second users (SUs), and the maximum tolerable interference constraints at primary users. However, the formulated optimization problem is non-convex and hard to tackle. By exploiting the classic semi-definite relaxation and successive convex approximation, we propose a penalty function-based algorithm to solve the non-convex problem. The convergence of the proposed algorithm is further proved. Finally, simulation results demonstrate that the non-linear EH model is able to strongly reflect the property of practical energy harvester and the performance gain of the proposed algorithm than the baseline scheme.