Distributed economic control strategy based on reinforcement pinning control for microgrids.

• Adjusting distributed generations output based on marginal cost consistency. • Integrating pinning control and Q -learning algorithm via parameter passing. • Simulations demonstrate that adopting the reinforcement pinning algorithm can reduce operating costs of the microgrid and improve system stability. Traditional droop control allocates distributed generation (DG) power based on capacity proportion, which leads to high system operating costs. To address this issue, this study proposes a distributed economic control strategy for microgrids based on reinforcement pinning (RP) control. To minimize operating costs, reinforcement learning (RL) continuously updates the Q -value table through reward feedback to obtain the optimal strategy. The optimal action determined by RL is set as the pinning reference value and transmitted to the pinning agent. Other agents achieve marginal cost consistency through the pinning information iteration matrix. To correct frequency deviations, proportional-integral (PI) control is used to ensure frequency stability for the system's operation. Simulations under different scenarios were conducted in MATLAB/Simulink. The results show that the proposed approach can coordinate the output of distributed power sources, reduce the operating costs of the microgrid, and maintain frequency stability. [ABSTRACT FROM AUTHOR]