Electricity cost optimization for Data Centers in smart grid environment

With the cloud computing development in recent years, electricity cost associated with Data Centers had been surging significantly. Hence, it is important for cloud service providers to reduce electricity cost as much as possible. In this paper, we conceive a strategy for reducing the electricity cost by utilizing energy storage facilities which exist in data centers in smart grid environment. The problem of electricity cost minimization in the data centers with energy storage facilities by considering the temporal variation of electricity prices is modeled as Markov decision process. And we propose a dynamic energy storage control strategy based on the Q-Learning algorithm which does not assume any a priori information on the Markov process governing the energy management system of the data centers. Finally, the proposed scheme is evaluated by numerical studies. The numerical results show that the proposed scheme can minimize the electricity cost of a data center.