Abstract/Summary

Fast-tracking a project involves carrying out sequential activities in parallel, partially overriding their original order of precedence, to reduce the overall project duration. The current predominant mathematical models of fast-tracking are based on the concepts of activity sensitivity, evolution, dependency and, sometimes, information exchange uncertainty, and aim to determine optimum activity overlaps. However, these models require some subjective inputs from the scheduler and most of them neglect the merge event bias. In this paper, a stochastic model for schedule fast-tracking is proposed. Relevant findings highlight the existence of a pseudo-physical barrier that suggests that the possibility of shortening a schedule by more than a quarter of its original duration is highly unlikely. The explicit non-linear relationship between cost and overlap has also been quantified for the first time. Finally, manual calculations using the new model are compared with results from a Genetic Algorithm through a case study.