Distributed Deep Reinforcement Learning for Intelligent Load Scheduling in Residential Smart Grids

The power consumption of households has been constantly growing over the years. To cope with this growth, intelligent management profile is necessary, such that can save electricity bills, and stress to grid during peak hours be reduced. However, implementing a method challenging due existence randomness in price appliances. address challenge, article, we employ model-free for households, which works limited information about uncertain factors. More specifically, interactions between modeled as noncooperative stochastic game, where viewed variable. search Nash equilibrium (NE) adopt based on distributed deep reinforcement learning. Also, proposed preserve privacy households. We then utilize real-world data from Pecan Street Inc., contains more than 1000 evaluate performance method. In average, results reveal achieve around 12% reduction peak-to-average ratio 11% load variance. With approach, operation cost