Dynamic Pricing Optimization for Load Aggregators in Virtual Power Plants Considering User Clustering and Response Behavior

Abstract

Load aggregators (LAs) can integrate end users with adjustable loads within a specific region and participate in electricity spot markets to mitigate the growing peak–valley imbalance in power demand. As an important enabler of demand-side flexibility and user participation, LAs play a critical role in enhancing the operational efficiency of new-type power systems. This study proposes a two-stage dynamic pricing framework for LAs within virtual power plant (VPP) environments. In the first stage, a K-means++ clustering algorithm is applied to classify end users according to their load characteristics and demand response capabilities. In the second stage, a Stackelberg game is formulated to characterize the hierarchical interaction between the LA (leader) and end users (followers), taking into account operational constraints and market uncertainties. Four simulation scenarios are designed to validate the proposed model. The results show that the optimized pricing strategy reduces electricity costs by 3.57% for non-residential users and 2.45% for residential users, while improving load smoothness and LA profitability. The proposed framework provides a practical and scalable approach for optimizing LA pricing strategies and offers valuable insights for enhancing user participation and renewable energy integration in real-time electricity markets.