A data-based model predictive control approach for enhancing energy efficiency in an air-cooled datacenter

This study explores the application of a databased Model Predictive Control (MPC) in optimizing data center cooling systems, emphasizing precise thermal regulation and energy savings. By integrating a dedicated neural network within the MPC framework, significant improvements are observed in both thermal management accuracy and energy efficiency. This hybrid strategy demonstrates potential for intelligent and sustainable data center operations, paving the way for enhanced management practices amidst escalating digital infrastructure demands.