Simulation-Driven Analysis and Modelling of Infrastructural Construction Works Performance

Abstract

The assessment of performance in infrastructural construction works is essential for improving efficiency and managing uncertainty in project delivery. This study proposes and empirically validates a simulation-driven framework for probabilistic modelling of construction duration based on statistically fitted productivity distributions. Sewerage construction works are analysed as a representative case of linear infrastructural projects characterised by repetitive production cycles and measurable daily outputs. Field data on daily work output, crew size, and working hours were statistically examined to identify variability patterns and dependency structures. After outlier assessment and goodness-of-fit testing at a significance level of α = 0.05, the Pearson Type III distribution provided the most consistent representation of daily productivity. The research tests the hypothesis that empirically fitted probability distributions of daily work output constitute a statistically valid basis for Monte Carlo simulation of project duration and deadline-compliance probabilities. Monte Carlo simulations performed for a 1,000 m work section (1,000–10,000 runs) produced stable and convergent results, yielding mean and median completion times of approximately 50.6 and 51 days, respectively. The probability of completion within 50 days was estimated at approximately 0.48, increasing to 0.75 within 51 days. The results confirm that distribution-based stochastic modelling provides a reliable and transferable framework for probabilistic assessment of performance in linear infrastructural construction works under uncertainty.