Joint Long-term Energy Efficiency Optimization in C-RAN with Hybrid Energy Supply

In energy harvesting cloud radio access network (C-RAN), energy cooperation technique can be utilized to improve system performance with the recent advances in smart grid. This paper investigates a long-term energy efficiency (EE) optimization problem in hybrid energy supplied C-RAN. In order to maximize the long-term utility and still maintain the network stability, a stochastic optimization problem is formulated. Particularly, both of the instantaneous quality of service (QoS) requirements and time-averaged limits are considered in the problem. By introducing Lyapunov framework together with the drift-plus-penalty method, we propose an EE-based power allocation and energy cooperation (EE-PAEC) scheme by jointly optimizing both of the sustainable energy and the grid power. The power allocation scheme contributes to optimal long-term EE with the deterministic upper bound of the data queue backlog, without requiring any prior knowledge of the dynamic system. Furthermore, theoretical analysis and simulation results demonstrate the performance advantage of energy cooperation, and show that the proposed scheme can provide a balanced trade-off between utility and stability.