Distributed Multi-Agent Energy Management Method for Microgrids via State-Based Potential Game

To address multi-agent self-interest and privacy protection requirements in microgrids, this paper proposes a distributed energy management method based on State-based potential game. First, wind turbine operators, photovoltaic operators, conventional generator operators, energy storage system operators and load aggregators are modeled as independent decision making agents, each with its own local profit and constraint model. Then, an opportunity cost correction is introduced to ensure forward-looking charge/discharge behavior under time-of-use pricing. Next, by incorporating over-limit estimators, the global power balance constraint is transformed into local penalty functions, achieving information decoupling and ensuring that the global system constraint is satisfied under the noncooperative game framework. Finally, a potential function is designed to ensure consistency with local objectives. All agents exchange only over limit estimators and converge to a pure strategy Nash equilibrium through gradient ascent iterations. Case study verifies the effectiveness of the proposed method.