Data-driven analysis and optimization strategies for power consumption in Electronic Gaming Machines

Abstract

Electronic Gaming Machines (EGMs) operate continuously in commercial environments, resulting in substantial long-term energy consumption. Despite their wide-spread deployment, there is limited empirical understanding of how different system configurations and operational conditions influence power usage at the system level. This thesis presents a data-driven case study aimed at characterizing the power and thermal behavior of an operational EGM and identifying opportunities for energy optimization. A non-intrusive measurement framework was developed to capture system-level power consumption and monitor thermal conditions at key locations. Experiments were conducted across variations in display brightness, sound levels, ventilation conditions, and representative system events. The collected dataset was processed and analyzed to evaluate how different factors influence overall power consumption. Based on the findings, the study identifies practical opportunities for reducing energy consumption through configuration-level adjustments and improved operational strategies. This work provides an empirical foundation for understanding energy behavior in continuously operating EGMs. The proposed framework is extensible to future work involving subsystem-level measurements, predictive modeling, and real-time optimization.