Multi-task deep learning economic dispatch of microgrids with electric vehicles and renewables
Seyed Morteza Ghorashi, Javad Khazaei, Shalinee Kishore · Sustainable Energy Grids and Networks · 2025
The increasing penetration of electric vehicles (EVs) and renewables in microgrids stimulates solving real-time economic dispatch (ED) that captures the stochastic nature of the problem. However, utilizing conventional or meta-heuristic methods to solve ED in real time is difficult and computationally costly, especially when addressing uncertainties of EVs and renewable energy sources. This paper proposes a multi-task deep learning approach to solve ED and separately learn the variability associated with EV and non-EV assets. The resulting case studies demonstrate the efficacy of the proposed data-driven model in solving real-time ED much faster and more scalable than numerical optimization and more accurately than conventional deep learning models.