abstract
- Global energy demand is increasing, and the existing buildings considerably contribute to this issue. Building energy retrofit as an effective method to improve energy efficiency is a proven yet challenging approach to reduce energy demand. In this respect, building energy retrofit requires assessment of a wide range of potential combinations of improvement measures. In addition, traditional optimization methods are computationally demanding and time-consuming, particularly for large buildings with many optimization parameters. This study aims to investigate the application of a state-of-the-art optimization algorithm, MEVO, to improve the building energy retrofit process. MEVO is a metamodel-based evolutionary optimizer that integrates machine learning and metaheuristic optimization with an active learning approach. The framework is developed by integrating Python as an optimization environment with EnergyPlus as an energy simulation engine to offer increased computational capabilities. This framework is applied to a four-story dormitory building on the University of Ottawa campus in Canada. MEVO was compared to the Bayesian optimization algorithm and demonstrated superior performance by achieving better optimization results in an equal number of function evaluations. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.