On the completion time of a project with random activity durations on a model of stochastic marked graphs

Abstract

Activities of a project often have random durations. This makes the estimation of the total project more difficult. Simulation, approximation and bouding techniques help the project manager in formulating these estimates. This paper focuses on formulating an upper bound on the project length based on the fact that an activity network can be represented as a stochastic marked graph. The bound, however, includes a part which requires that the distributions of the activity durations are fully specified. As many times, only incomplete information is available on these durations like the interval, the mean, the mode or the standard deviation, this research investigates how the bound should be computed including only incomplete information.