A Demand response regulation method considering dynamic user propensities

Abstract To address challenges in the intricate governance of customer resources within power grids heavily integrated with renewable sources, which lead to high demand response costs and low user satisfaction, a demand response regulation method considering dynamic user propensities is proposed. Firstly, the user-side aggregator (UA) is defined to participate in demand response. A model of dynamic user propensities is established to evaluate user satisfaction. Secondly, the optimization problem is set as minimizing the daily operation cost of the UA’s demand response. Finally, a demand response regulation optimization algorithm, utilizing enhanced reinforcement learning (RL), is introduced to address the optimization challenge. A soft update is performed by adjusting the network learning rate based on user satisfaction, allowing for greater flexibility. The study’s simulations indicate that the suggested method significantly cuts the daily operational expenses of demand response while satisfying users’ evolving economic and comfort needs.