Applying Reinforcement Learning Method for Real-time Energy Management

Today energy management and optimization is a key factor to control the whole production cycle, distribution and energy consumption. Electrical energy consumption optimization involves proper modeling and prediction. Preservation of energy-efficient resources and proper consumption management is one of the most important challenges in all countries of the world. In this study, we present a decision support system for managing energy by reinforcement learning. First, a set of different energy uncertain consumption data and adopted decisions were considered in the form of fuzzy. Then, the prediction of consumption was done by the Q-learning algorithm, which is a solution to the Markov decision problem. Then the rules are presented to describe what the system implies. The proposed method is capable of working in real-time approach and handle the consumption fluctuations in learning and predicting process.