Integrated Bayesian-Monte Carlo Based Probabilistic Risk Modeling of Cost Overruns in Infrastructure Projects

Abstract

Cost overruns in road construction projects significantly challenge project efficiency and financial sustainability. This study develops an integrated Bayesian-Monte Carlo framework to identify and quantify critical cost overrun factors in Pakistan's road infrastructure projects. Through literature review and expert consultation, 25 potential causes were systematically reduced to 10 critical factors using expert elicitation and criticality scoring. Bayesian analysis calculated posterior probabilities by combining expert judgment with conditional relationships, while Monte Carlo simulation quantified uncertainties and provided probabilistic ranges for each factor. The framework reveals design error changes as the most critical factor (mean: 61.5%, maximum: 93.3%), followed by variation orders (58.9%) and inaccurate estimates (57.8%). Secondary contributors include land acquisition issues (54.5%) and schedule delays (52.3%). This integrated approach enables evidence-based risk prioritization, replacing arbitrary contingency planning with data-driven decision making. The methodology provides project managers with actionable insights for targeted risk mitigation and optimal resource allocation in resource-constrained environments.