Uncertainty-Guided Optimal Scheduling of Automated Energy Internet

This article tries to propose a novel uncertainty-guided day-ahead scheduling framework for the microgrid energy Internet incorporating the renewable energy sources, battery storage, and electric vehicles. The proposed model makes the use of the prediction interval concept to capture the uncertainty effects in the bandwidth and create an optimal interval for the output cost function. To this end, a deep learning approach based on the generative adversarial models (GAMs) is designed to not only have a prediction interval with high confidence level but also make the least bandwidth with the maximum possible information. Moreover, an evolving approach based on bat algorithm is proposed to help to train the GAM optimally. Finally, the model performance and quality are assessed using the IEEE standard test system.