Data-Driven Decision Framework for Modeling and Quantification of Uncertainty and Variability in Construction Project Scheduling

The success of an infrastructure project is often judged by its timely completion within budget and quality standards. Most construction projects experience delays due to poor project planning, budget conflicts, project complexity, weather and natural hazards, injuries and safety issues, and management issues. While the construction industry primarily uses network-based scheduling approaches that rely on the existence of a conventional critical path estimated using deterministic activity times. These approaches do not account for uncertainties and variabilities associated with activities and ignore the possibility of alternate paths becoming critical due to changes incurred in activity times. This research uses a simulation-based approach to demonstrate the impact of the stochastic nature of activity times and to quantify the impact of variation in activity times on project completion. The study presents a novel data-driven decision framework to determine a robust project planning horizon that effectively accommodates uncertainties and variabilities in construction activities. This refined approach was found to be effective for initial project scheduling and for estimating the time extensions required if the project incurs any delay during execution. The results clearly show that minor changes in activity times can lead to the emergence of alternate critical paths and can reduce the probability of timely completion of construction projects. The study recommends the use of average completion times (Cmax) estimated using Cmax of all possible critical paths in the network, which can provide an accurate estimate of project completion time. In addition, a path with the highest variance can be different from the critical path estimated using deterministic activity times. Hence if any delays incur during execution, additional resources need to be invested on paths with higher variance to bring project completion time back to acceptable levels. The study also proposes a graphical approach to estimate realistic completion times, enhancing the probability of achieving timely project delivery amidst uncertainties related to risk, reliability, and natural hazards.