Online energy management strategies for base stations powered by the smart grid

We consider the problem of minimizing the electricity bill for a cellular base station powered by the smart grid and locally harvested renewable energy. We consider hourly-varying electricity prices made known one day ahead to the base station. We assume that the base station is equipped with a finite-capacity battery. We ensure that the instantaneous energy demand of the base station is satisfied and the constraints imposed by the battery are observed at any point in time. We propose several online energy management strategies that require only causal knowledge of the renewable energy generation and the power consumption profiles. We benchmark our proposed strategies against the optimal energy management policy which assumes perfect knowledge of all system parameters, e.g., base station energy usage and renewable energy generation, both in the past and the future. Simulation results show that the performance of our proposed online strategy deviates from the optimal by 2% at most.