New competitive results for the stochastic resource-constrained project scheduling problem: exploring the benefits of pre-processing

We study the resource-constrained project scheduling problem with stochastic activity durations. We introduce a new class of scheduling policies for solving this problem, which make a number of a-priori sequencing decisions in a pre-processing phase while the remaining decisions are made dynamically during project execution. The pre-processing decisions entail the addition of extra precedence constraints to the scheduling instance, hereby resolving some potential resource conflicts. We obtain new competitive results for expected-makespan minimization on representative datasets, which are significantly better than those obtained by the existing algorithms when the variability in the activity durations is medium to high.