Machine Learning for Energy Optimization in Computing Systems
Monika Bansal, Nishi Prakash Jain, Nitish Pathak, Neelam Sharma · 2025
The development of innovative technologies has resulted in an increase in the requirement for energy-efficient and high-performance computing systems. One of the 17 UN Sustainable Development Goals, which are international development goals established by the United Nations in 2000, emphasizes energy optimization. These goals are part of a broader agenda to be implemented by 2030 and seek to promote development and equity across nations that require virtually all sectors of society, including networking, software, hardware, and other computer system components, to be energy-efficient. As a part of continuous overall improvement and productivity enhancement, machine learning offers high levels of automation, data-driven decision-making, and predictive analysis, for example, in the automation of processes. According to the latest research, the importance of machine learning in reducing electricity expenditures and consumption can hardly be overstated, as it aids in engineering-controlled network load distribution, task management, supply of resources, and reduction of dynamic hardware components. This contributes to the protection of the environment and benefits the economy at the same time. This research investigates advancements in machine learning to boost performance and reduce energy consumption in computing applications such as data centers, cloud facilities, smart grids, and mobile devices. First, energy-efficient computing systems are introduced, and the significance of such systems and the techniques of machine learning in such systems are explained. Then, applications of computational systems are considered from the point of view of specific machine learning applications for energy efficiency at appropriate computational tiers. It reviews past developments, presents challenges and research needs, and suggests possibilities for future developments. Finally, the chapter outlines the scope of green artificial intelligence (Green AI), an emerging direction in energy conservation.