A Hybrid Project Risk Intelligence Algorithm for Predicting Cost Overruns and Schedule Delays in Infrastructure Projects: Comparative Evaluation with Monte Carlo Simulation and Critical Path-Based Risk Assessment

Abstract

Cost overruns and schedule delays remain persistent weaknesses in infrastructure project delivery, especially in transport, energy, water, public building, and large civil works programmes. Traditional project-risk tools such as Monte Carlo Simulation and Critical Path Method-based risk assessment provide useful analytical support, but their performance is limited when project behaviour is nonlinear, data are incomplete, and cost-schedule interactions evolve during execution. This paper proposes a Hybrid Project Risk Intelligence Algorithm, named HyPRIA, for predicting cost overrun and schedule delay risk in infrastructure projects. HyPRIA integrates earned value indicators, critical path metrics, risk-register scores, procurement variables, design-change indicators, Monte Carlo uncertainty features, and ensemble machine learning into a unified predictive framework. The proposed model is comparatively evaluated against Monte Carlo Simulation and Critical PathBased Risk Assessment using a controlled illustrative dataset representing 420 infrastructure projects across road, bridge, rail, energy, public building, and water infrastructure categories. Results indicate that HyPRIA achieves stronger predictive performance than the baseline models, with lower mean absolute error, lower root mean square error, higher coefficient of determination, stronger delay-classification performance, and better probability calibration. For cost overrun prediction, HyPRIA achieved an RMSE of 3.91 percentage points compared with 8.94 for Monte Carlo Simulation and 10.12 for CPMbased risk assessment. For schedule delay prediction, HyPRIA achieved an RMSE of 4.20 percentage points compared with 9.85 and 11.40 for the two baselines. Explainability analysis shows that float consumption ratio, procurement delay index, design-change frequency, cost performance index, schedule performance index, risk exposure score, and contractor productivity deviation are dominant predictors. The study contributes a technically interpretable hybrid risk-intelligence framework for early warning, project governance, and proactive mitigation of cost and schedule failures in infrastructure delivery.