A Systematic Taxonomic Review of Risk Modelling and Assessment Methods in Construction Projects (1990–2025)

Abstract

This study presents a systematic taxonomic review of risk modelling and assessment methods in construction projects over the past 35 years (1990–2025). Through a structured four-stage process, 91 peer-reviewed articles from 15 leading journals were analysed. The taxonomic approach enabled the classification and mapping of methods according to chronological evolution, study type, authorship patterns, and focus areas, while thematic analysis was employed to synthesise key themes, trends, and research gaps. The review examines publication trends, geographical distribution of research contributions, and methodological developments. The findings reveal that the probability-impact (P-I) model remains the dominant approach, despite its well-documented limitations in capturing risk interdependencies and their cascading effects on project quality and overall performance. Fuzzy Set Theory (FST), Analytic Hierarchy Process (AHP), and Monte Carlo Simulation (MCS) emerged as the most frequently adopted techniques. The analysis demonstrates a clear evolution in the field: from predominantly basic probabilistic methods in the 1990s to increasingly sophisticated hybrid, fuzzy logic-based, and AI-enhanced approaches after 2010. Notwithstanding these advancements, significant gaps persist, particularly the lack of integrated frameworks capable of simultaneously addressing risks across multiple project objectives—cost, time, quality, and performance. This review synthesises the state of knowledge in the field, identifies persistent theoretical and practical shortcomings, and offers a comprehensive roadmap for future research. Key directions include the development of machine learning applications, dynamic modelling techniques, and holistic multi-objective risk assessment frameworks to better align risk management theory with the complex realities of modern construction projects.