Multi-Agent H-MG Energy Management Based on Continuous Quadratic Real-Time Q-Learning Method

The increasing popularity of connected electric vehicles (EVs) and renewable energy sources requires the design of advanced energy management systems (EMS) to optimize energy use and distribution in home microgrids (H-MGs). This study focuses on energy management in H-MGs and analyzes its impact on grid. In this research, it has been shown that a cooperative multi-agent EMS system managed by continuous and real-time quadratic Q-learning method can reduce the consumption of electrical loads by 0.9% and thermal loads by 24%. Compared to the fuzzy logic control (FLC) method, electricity consumption can be reduced by 21% and gas consumption by 17%, as a result of which the cost is reduced by 18%.