Proactive Project Scheduling with Time-dependent Workability Uncertainty

Proactive scheduling can effectively handle activity duration uncertainty in real-world projects, by generating a baseline solution according to a prior stochastic knowledge. However, most of the previous approaches cannot deal with the activity duration uncertainty caused by time-dependent workability uncertainty. In this paper, we aim at finding a partialorder schedule (POS) that produces the minimum expected makespan on a given probability model of workability uncertainty. Since this is a hard discrete stochastic optimization problem, we propose an approximation approach based on Sample Average Approximation (SAA), and develop a branch-and-bound algorithm to optimally solve the SAA problem. Empirical results on benchmark problem instances and real-world distribution data show that our approach outperforms the best general-purpose POS generation approaches that do not exploit the stochastic knowledge.