Proactive policies for the stochastic resource-constrained project scheduling problem

The resource-constrained project scheduling problem involves the determination of a schedule of the project activities, satisfying the precedence and resource constraints while minimizing the project duration. In practice, activity durations may be subject to variability. We propose a stochastic methodology for the determination of a project execution policy and a vector of predictive activity starting times with the objective of minimizing a cost function that consists of the weighted expected activity starting time deviations and the penalties or bonuses associated with late or early project completion. In a computational experiment, we show that our procedure greatly outperforms existing algorithms described in the literature.

► In this paper we study the RCPSP when activity durations are uncertain.
► We consider positive and negative deviations from the predictive activity starting times of inflexible activities.
► Costs are also associated with exceeding the project due date and a bonus is awarded when finishing early.
► We derive optimal predictive starting times based on the newsvendor problem.
► Our procedure is vastly superior than the current state-of-the-art for this problem.