Multi-Objective Optimal Scheduling of Energy and Computational Workload in Green Data Centers Considering Waste Heat Recovery

Driven by the growth of the digital economy, the demand for computational performance of data centers (DCs) is growing exponentially, which leads to huge energy consumption in DCs. As such, this paper proposes a multi-objective optimal scheduling method for green DCs to jointly manage energy supply and computational workload. Firstly, an electricity-heat-computation coupling system model is constructed, which considers renewable energy sources (RESs), waste heat recovery and the temporal flexibility of DC workload. Then, a multi-objective optimal scheduling model is formulated with the comprehensive consideration of economic, environmental and energy efficiency indicators. Further, the power usage effectiveness (PUE) is modified to reformulate the objective as a linear function. Finally, the proposed solution is assessed extensively through a case study, and the numerical results confirm the efficacy of this method in reducing economic cost, carbon emissions and PUE.