A Machine Learning-Based Energy Management Agent for Fine Dust Concentration Control in Railway Stations

This study developed a reinforcement learning-based energy management agent that controls the fine dust concentration by controlling facilities such as blowers and air conditioners to efficiently manage in station. To this end, we formulated an optimization problem based on Markov decision-making process model for predicting of station training artificial neural network (ANN) supervised learning develop transfer function. In addition prediction model, optimal policy blower conditioner according current state was obtained ANN which Deep Q-Network (DQN) algorithm applied. case study, it is confirmed DQN predictive were trained actual data Nam-Gwangju Station converge policy. The comparison between proposed method conventional shows can use less power consumption but achieved better performance reducing than method. addition, increasing value ratio represents compensation due reduction, learned more reduction conditioner.