Designing an Intelligent Decision-Support Model for Risk and Delay Management in Industrial Projects Using an Integration of Monte Carlo Simulation and Machine Learning in Dynamic Uncertainty Environments

Abstract

In recent years, industrial projects have faced increasing complexity, uncertainty, and interdependence among project activities. According to international project management reports, approximately 60% to 80% of large-scale industrial projects experience schedule delays, and nearly 45% encounter cost overruns of more than 20% compared with the initial budget. The objective of this study is to design an intelligent decision-support model for predicting and managing risk and delay in industrial projects by integrating Monte Carlo simulation and machine learning algorithms. In this study, performance data from previous projects, including activity schedules, resource consumption levels, cost variations, and risk indicators, are first collected. Then, using Monte Carlo simulation, more than 10,000 probabilistic project execution scenarios are generated to model uncertainty in time and cost from a probabilistic perspective. Subsequently, machine learning algorithms, such as Random Forest and Gradient Boosting, are applied to identify delay patterns and predict schedule deviation with a targeted accuracy of approximately 85% to 92%. The expected results indicate that the proposed hybrid model can improve the accuracy of project delay prediction by up to 30% compared with traditional project planning methods, including CPM/PERT, and reduce the mean error in estimating project completion time from approximately 18% to less than 10%. Furthermore, the model is capable of identifying critical risks at least 2 to 4 weeks earlier than conventional methods, which can play an important role in reducing additional costs and improving resource productivity.