Correcting Monte Carlo Bias In Tunnel Scheduling: A Comparative TD–Des Framework for Hybrid NATM–TBM Projects

This research proposes a corrected Monte Carlo (MC) framework to examine structural bias in probabilistic tunnel scheduling. The framework integrates parallel multi-front excavation modelling, One-At-a-Time sensitivity analysis, and an outer-loop Monte Carlo correction that explicitly captures parameter-level uncertainty. Time-driven and discrete event simulation approaches are evaluated using a common dataset and validated against field data from the Eurasia Tunnel project in Istanbul. Results show strong agreement with documented tunnel boring machine performance, with advance-rate deviation below 0.4%. Parallel multi-front execution reduces the predicted P50 project duration by 64.3% and cost by 14.9% relative to a fully sequential baseline, while a combined resource-optimisation scenario (increased crew size, equipment, and efficiency) yields a further 22.8–22.9% schedule improvement that brackets the project’s documented 17.1% early completion.