"An Integrated Demand Response Optimization Model Considering Social Learning of Users Under Partially Anomalous Information"

"Integrated Demand Response (IDR) is an effective means to ensure supply\u2013demand balance in modern power systems. However, existing IDR models fail to consider the social learning behavior of users and struggle to accurately estimate user parameters under partial anomalous information. To address these issues, this paper incorporates observational learning and word-of-mouth learning into user modelling and integrates L2 regularization into parameter estimation. First, an improved model representing users\u2019 observational learning behavior is constructed based on peer effect theory. Second, the evaluation metric tailored to IDR services is proposed, and a quantitative model for users\u2019 word-of-mouth learning behavior is developed. In addition, an improved parameter estimation method is proposed by integrating L2 regularization into the conventional maximum likelihood estimation to address partially anomalous information. Simulation results demonstrate that the proposed model not only enhances the effectiveness and economic efficiency of incentive strategies for multi-energy service providers (MESPs) and reduces supply\u2013demand imbalance risks, but also improves user benefits in participating in IDR. Furthermore, the proposed parameter estimator enhances the precision of parameter estimation under partial anomalous information, compared with traditional estimators."