Sustainable selection of construction materials using machine learning: A review

Abstract

Global environmental challenges predominantly originate from the construction industry, with material selection becoming critical for sustainability. Due to the complexity of maintaining a balance of economic, performance and environmental criteria, traditional methods have become insufficient for material selection. This chapter explores the application of machine learning (ML) in optimising sustainable construction material selection. ML techniques, such as supervised and unsupervised learning, predictive modelling, and multi-criteria decision-making, are examined for their ability to evaluate material properties and lifecycle impacts. It highlights the strength of ML models for predicting material performance, optimising the selection process and minimising carbon footprints, with an emphasis on economic and environmental factors. While ML offers significant potential, challenges such as data availability, alignment with industry standards and model interpretability need attention to realise its benefits. Advanced methods such as reinforcement learning, deep learning and integrating internet of things (IoT) sensors in real-time material selection are identified as key areas to be explored. ML helps in efficient decision-making processes, thus making it a transformative approach to achieving sustainable construction practices.