Machine Learning-Optimized Concrete Mix Design System For Sustainable High-Rise Construction
DOI:
https://doi.org/10.64252/8q081y93Keywords:
Machine learning, concrete mix design, sustainable construction, high-rise buildings, gradient boosting machines, genetic algorithms, multi-objective optimization.Abstract
This research introduces a machine learning approach to help improve sustainability in the construction of high-rise buildings. The idea was to merge Gradient Boosting Machines with Genetic Algorithms to build a framework that helps predict and reduce cement amounts, all while the concrete meets the 90-day compressive strength standard. With Python, using XGBoost and DEAP libraries, the system yields an R² score of 0.92, indicating effective accuracy. Because the mix design uses fewer cement resources, it also leads to a 12% drop in CO₂ emissions. When compared to other methods, this new approach performs better at achieving good designs and addressing problems in making construction sustainable. According to the results, the eco-friendly compositions have promising potential for industrial adoption, as their durability and strength are not affected. In the next phase, we will add in extra material parameters and live building updates to further improve the outcomes of our optimization.
