Stochastic project scheduling based on time lag

The purpose of this paper is the resource constrained scheduling of stochastic complex projects. Projects are modeled using serial/parallel graphs where the duration of activities are defined by random variables with piecewise polynomial distribution functions. In order to simplify the approach, the different activities share only one class of resources available with a limited capacity. We want to estimate the probability that a resource conflict occurs at a time t during the project execution. The scheduling problem to solve is then a two level problem: firstly it is a minimisation of the average duration of global project and secondly we have to guarantee that the conflict probability is below a given threshold