Optimizing support vector regression using grey wolf optimizer for enhancing energy efficiency and building prototype architecture

In this era of increasing energy demand, optimizing energy consumption in building systems is critical for enhancing sustainability and operational efficiency. Residential buildings consume significant energy, with heating load (HL) and cooling load (CL) being major contributors. While previous research explored modeling techniques, there's a need for more precise and optimized approaches. It has been observed that accurately predicting heating and cooling load can help in designing better building prototypes with improved energy efficiency. Researchers have explored multiple computational intelligence approaches that have improved the prediction of energy consumption of buildings. However, the applied approaches have some limitations. To improve the prediction of HL and CL of buildings and to overcome the limitations of the existing approaches, this study employs a support vector regression (SVR) fine-tuned with Grey Wolf Optimization (GWO) for enhancing energy efficiency and building prototype architecture by predicting HL and CL. To establish the supremacy of the proposed approach experiments are conducted on a benchmark dataset and performance is compared with four prominent machine learning (ML) algorithms namely linear regression, decision tree, random forest, and deep neural networks to predict HL and CL. The results demonstrate that SVR-GWO outperforms the others, achieving an r-score of 99.77% for HL and 97.90% for CL. Additionally, various metrics like Mean Absolute Error (MAE), Mean Square Error (MSE), and Root Mean Square Error (RMSE) are computed to evaluate model performance. The HL model achieved MAE, MSE, and RMSE scores of 0.3422, 0.2354, and 0.4852, respectively. Similarly, the CL model achieved scores of 0.8387, 1.9426, and 1.3937, respectively. The results show better performance of the proposed model in comparison to the ML models.