Research Progress on Load Forecasting and Scheduling Optimization for New Power Systems

With the gradual diversification of energy demand and consumption patterns, new types of loads—such as integrated energy systems and electric vehicles—are emerging. The characteristics of electricity consumption behavior and load components are becoming increasingly complex, reducing load forecasting accuracy in new power systems, and increasing the difficulty of grid load dispatching optimization. This paper first focuses on load forecasting in new power systems, analyzing classical methods like regression analysis, gray forecasting, and time series analysis, as well as artificial intelligence approaches such as artificial neural networks and wavelet analysis. It then examines modern ensemble forecasting methods for multi‐load integration scenarios, which address traditional methods' limitations in handling nonlinear and nonstationary loads while avoiding local optimization or overfitting in simple machine learning. Finally, it summarizes research on load‐side resource allocation and scheduling control using mathematical programming and genetic algorithms, with in‐depth discussion on synergistic optimization of forecasting and dispatching. Furthermore, this review maps methods to specific application scenarios, offering practical guidance for researchers and practitioners.