Distributed Online Energy Management for Data Centers and Electric Vehicles in Smart Grid

To avoid an additional large demand charge introduced by the adoption of electric vehicles (EVs), we investigate a joint energy management problem for geographically distributed data centers (DCs) and EVs of the employees in this paper. Specifically, we intend to minimize the long-term total cost of DCs and EVs by jointly scheduling DC workloads and EV charging demands, without violating heterogeneous EV charging requirements and the given power limits in all geographical locations. We first formulate a stochastic programming problem with the considerations of uncertainties in DC workloads, electricity prices, and EV energy demands. Since these uncertain system parameters are time-varying and the size of the formulated problem is large, we then propose a distributed online algorithm to solve the formulated problem based on Lyapunov optimization technique and a variant of alternating direction method of multipliers. Finally, extensive simulation results show the effectiveness of the proposed algorithm.