Research work · 2025

Optimization of Smart Communities Based on Artificial Intelligence

Bárbara Morales Díaz, Àlex Pujol Garcia, Regina Enrich Sard, Jose María Santos, Juan Trullos

Frontiers in Artificial Intelligence and Applications

Abstract

Energy communities face the challenge of efficiently managing the energy generated in order to ensure an equitable distribution among their members. In this context, this work presents an approach based on deep learning techniques and optimization algorithms to improve the allocation of energy by forecasting the annual consumption needs of the members of the community. The data flow includes the clustering of users according to their consumption profile as well as other characteristics, the forecasting of hourly energy consumption and production for the next 12 months and the calculation of the optimal energy share to be distributed among partners in an optimal way for members according to their profile. For the forecasting module deep learning technologies were explored as well as classical machine learning algorithms. Finally, the optimization algorithm combines metaheuristic algorithms with several post-process refinements. The goal of this approach is to facilitate decision-making in community management as well as in the evaluation of new members.

DOI: 10.3233/faia250612