Resource Allocation Policies for Hybrid Power-Grid and Harvested Energy Communication Systems

This work studies resource allocation policies for a multi-antenna access point that is powered by a combination of harvested energy and the power grid, communicating with multiple single-antenna users. We propose a non-convex problem to directly solve the throughput maximization problem. Though the problem is challenging to solve, we first propose an iterative algorithm based on the first-order Taylor expansion and block coordinate descent for the scenario that full channel state information (CSI) and energy arrival information (EAI) are assumed to be known. Then, inspired by this scenario, we study a case in which statistical CSI and EAI are only required. Simulation results demonstrate that the energy-performance trade-off as well as the performance of the statistical case is comparable to the full CSI and EAI scenario, which supports the practical aspect of the proposed policies.