Abstract

We introduce a simulation and optimization framework for stochastic resource-constrained project scheduling problems. The stochasticity is represented by modeling job durations as continuous random variables without time discretization, while fluctuations are captured by simulating a sufficiently large number of realizations. The objective is to analyze the relationship between variability in the input parameters and the resulting makespan in order to enable a priori estimates. Estimation approaches derived from simulations of small instances can then be extrapolated to problems involving a larger number of jobs or more complex characteristics.