Researchers from the University of Eastern Finland introduced a hybrid evolutionary-based algorithm, GA-SHADE.

GA-SHADE is designed to optimize machine learning models and feature selection for predicting energy consumption in district heating systems.

The algorithm supports efficient energy production and distribution in urban areas.

GA-SHADE can identify simplified ML models with good prediction performance and adapt without fine-tuning.

It handles varying numbers of features and hyperparameters effectively.

Support Vector Regression (SVR) and Neural Networks (NN) were identified as the best performing models for this purpose.

The study's findings contribute to the development of accurate predictive models for energy consumption in district heating networks.

The research underscores the role of advanced methodologies in enhancing sustainability and reducing carbon emissions in building operations.