Project scheduling via stochastic programming

Abstract

If for a project (described by a non-empty set of activities, a relation on this set of activities the transitive closure of which is a strict order, and activity-completion-times assigned to the single activities) the activity-completion-times are assumed to be random variables a two-stage stochastic programming approach can be used for a cost-oriented project scheduling model. Completion-time estimates for the activity-completion-times are computed in such a way that, in order to meet a prescribed time-constraint for the project-completion-time, the expected costs for performing the activities according to the computed time-schedule are minimized. An example is included for illustration.