Optimizing Indoor Air Quality and Energy Use in High-Density Asian Apartments Using Machine Learning and Metaheuristic Algorithms

The present paper explores the growing problems of poor indoor air quality (IAQ) and elevated energy consumption recorded in densely populated Asian apartment blocks. This research aims to propose a novel hybrid approach that can optimize both IAQ and energy performance by utilizing sophisticated artificial intelligence methods. For this purpose, five machine learning models, Artificial Neural Network (ANN), Support Vector Machine (SVM), Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Convolutional Neural Network (CNN), were combined with ten popular metaheuristic algorithms: Genetic Algorithm, Particle Swarm Optimization (PSO), Grey Wolf Optimizer (GWO), Harmony Search (HS), Firefly Algorithm (FA), among others. As a result, a significant improvement in the models' performance, achieved due to efficient hyperparameter tuning and expressed as lower MSE and MAE rates and higher R² values, was identified. More precisely, the best performances were provided by XGBoost and CNN models. Besides, it should be noted that the correlation between the environmental factors (CO₂ concentration, temperature, etc.), analyzed within the present research, and energy consumption appeared very high. Moreover, sensitivity analysis proved the applicability and robustness of the proposed models to various noise levels. Nonetheless, the research has several limitations, which are mainly associated with the use of both simulated and sensory data, along with the selection of a limited number of environmental factors for consideration. In practice, the framework developed within this research may prove useful as an efficient tool in decision-making for optimizing energy consumption in smart buildings, especially in the context of modern architecture and engineering. Socially, the research is relevant to the improvement of human well-being and the development of sustainable cities.