Feasibility of machine learning application in pavement life cycle assessment: A review

The need to mitigate the environmental impacts of pavement systems has increased interest in life cycle assessment (LCA), but its implementation often faces challenges, such as data uncertainties, inconsistent impact methods, and limited decision-support capabilities. This review explores the utilization of machine learning (ML) to address these challenges and enhance LCA workflows. This review addresses the current practices and challenges in pavement LCA by structuring it around its four phases, i.e., goal and scope definition, inventory analysis, impact assessment, and interpretation. Diverse applications of data-driven ML techniques in pavement systems and LCA are highlighted. Review indicates that ML can enhance pavement LCA by predicting context-specific inventory data, clustering diverse datasets to detect inconsistencies, and simulating different allocation scenarios. Moreover, multiple impact categories forecasting seems possible with ML-based inventory analysis. ML-based visualisations, such as decision trees, can clarify variables’ contributions to environmental outcomes. ML can also support sensitivity and uncertainty analyses to strengthen decision-making.