Machine Learning based Energy Management System for Smart Buildings

Abstract

Research and development on energy management in smart buildings has been of utmost importance in recent years
due to the growing need for efficient and sustainable urban settings. By utilizing reinforcement learning to
dynamically choose the optimal ventilation equipment (air conditioners, fans, and windows) in response to variable
environmental conditions like humidity, temperature, and wind speed, this work fills a gap in the literature on
energy management and improves smart building energy management. In order to train the ML model, we used
reinforcement learning (RL) with a fuzzy control scoring system that took into account the device's power cost and
the consequent user satisfaction as a function of weather conditions (temperature, humidity, and wind speed). In
addition, case studies demonstrating its use in both normal and exceptional situations allowed an integrated
simulation platform to validate the model's decision-making process.