Key points are not available for this paper at this time.
As global energy consumption continues to rise, Data Centers have emerged as a critical driver of energy demand. While technological innovations have improved energy efficiency, the increasing reliance on digital services requires sustainable energy management solutions. The motivation for this study is to support the efficient use of energy in Data Centers and help operators make informed, data-driven decisions, promoting more engagement between academic research and industrial practice. This paper presents a comprehensive literature review on forecasting and optimizing the energy consumption and supply of Data Centers. It examines state-of-the-art machine learning models for energy prediction, including time-series forecasting and deep learning techniques, alongside various optimization strategies aimed at improving energy efficiency and maximizing renewable energy usage. A key focus is bridging the gap between prediction and optimization, ensuring that Data Centers can dynamically adapt to fluctuating energy availability while maintaining operational efficiency and reliability. By reviewing existing methodologies, this study offers valuable insights for industry professionals seeking to enhance sustainability, reduce operational costs, and lower carbon emissions. As a part of this review, a compilation of available R package implementations of the machine learning models is also provided, offering practical guidance for reproducibility. The findings contribute to the development of data-driven decision-making tools that support the transition toward 24/7 Carbon-Free Energy in the digital infrastructure sector.
Ferreira et al. (Tue,) studied this question.