Assessment of global sensitivity analysis methods for project scheduling

Autores UPV
Año
Revista COMPUTERS & INDUSTRIAL ENGINEERING

Abstract

Product development requires scheduling that considers the interdependence between activities. The definition of the interdependencies and duration of activities, communication times and the level of overlap between activities is needed for project scheduling. However, these parameters have epistemic uncertainties that can affect project scheduling. In this work, different global sensitivity analysis techniques were applied to identify the parameters that had the greatest effect on project scheduling. It was concluded that standardized regression coefficients as well as the Morris and Sobol¿¿Jansen methods were the most appropriate. It was also found that global sensitivity analysis can help to focus resources based on the definitions and control the uncertainty of key activities. Furthermore, it was concluded that control of the uncertainty of key activities reduces the uncertainty and duration of projects.