Electric vehicles and smart grids

Smart energy management is becoming one of the major challenges of the transition to electric mobility, especially when multiple vehicles, users, and networks must coordinate in real time without compromising the system's efficiency.

In this context, the seminar “Deep Reinforcement Learning in a Multi-Agent System for Defining Electric Vehicle Load Management Actions” was held, organized by the DemIA International Chair in collaboration with the University of Salamanca’s doctoral program in Artificial Intelligence, as a forum for analyzing the role of artificial intelligence in optimizing energy networks.

The session explored the use of deep reinforcement learning applied to multi-agent systems, where different nodes autonomously learn to make coordinated decisions to manage the load of electric vehicles. This approach allows for the dynamic adjustment of energy consumption based on grid availability and user behavior, reducing demand peaks and improving system stability.

The presentation was given by Carlos Ernani da Veiga, M.Eng., a researcher in the Department of Electrical Engineering at the Federal Institute of Santa Catarina (IFSC), who has extensive experience in developing solutions based on smart systems for power grids and energy optimization.

During his presentation, he introduced a model in which each agent incorporates contextual information from the environment to learn adaptive action policies. This approach allows for the integration of heterogeneous usage patterns and the anticipation of charging needs, achieving more efficient coordination between users and infrastructure.