DNN-Based EV Scheduling Learning for Transactive Control Framework

Research output: Chapter in Book/Report/Conference proceedingBook chapterEducationpeer-review


This study employs deep neural networks (DNNs) to determine a suitable charging schedule for electric vehicles (EVs) while avoiding power losses in the distribution network and ensuring that voltage constraints are not violated. A network-constrained transactive control framework, including aggregators, distribution system operators (DSOs), and a price coordinator, was established to generate a data set for learning. The price coordinator played a crucial role in promoting interaction between DSOs and aggregators and promoting convergence. By repeating the process and adjusting the inputs, a data set was produced that could be utilized for learning. To assess the performance of DNN, the results were evaluated against system constraints and all samples showed satisfactory results without constraint violations, demonstrating that a well-constructed data set enables DNNs to effectively learn complex problems with guaranteed convergence in a limited time frame for transactive control problems.
Original languageEnglish
Title of host publicationMicrogrids: Theory and Practice
EditorsPeng Zhang
PublisherJohn Wiley & Sons Ltd
Publication date2024
ISBN (Print)9781119890850
ISBN (Electronic)9781119890881
Publication statusPublished - 2024


Dive into the research topics of 'DNN-Based EV Scheduling Learning for Transactive Control Framework'. Together they form a unique fingerprint.

Cite this