Fuzzy Q-Learning based energy management of small cells powered by the smart grid

摘要

With the rapid increase of mobile data demand, Mobile Network Operators (MNOs) have started to pay close attention to wireless network energy consumption and CO2 emissions. In the current context of energy transition, Renewable Energy (RE) represents a great potential for MNOs to use environment-friendly power supply and reduce the energy expenses. Moreover, the Smart Grid (SG) offers important services in terms of Demand Side Management and decentralized production, which make rethinking the power usage within mobile networks a necessity. In this paper, we propose a fuzzy Q-Learning based energy controller for a small cell powered by local RE, local storage, and the SG to simultaneously minimize electricity expenditures of the MNOs and enhance the life span of the storage device. Simulation results show that the proposed solution achieves important cost reduction with respect to simpler approaches and performs very closely to the ideal strategy based on a perfect knowledge of the stochastic variables.