A Bayesian Network model to measure project manager's confidence on project execution

Abstract

Bayesian Networks are a well-recognized decision support tool for a wide range of situations that involve uncertainty and probabilistic reasoning. Projects are not an exception to these situations. They are usually one-off experiences where many data are incomplete, suffer from imprecision and accuracy, and estimations are conditionally dependent on assumptions that are a major source of uncertainty. This uncertainty is present during the different stages of a project where complex causal relationships are involved between uncertainty and project performance. In this complex and uncertain environment many project managers are faced with the question, will the project finish successfully? This paper introduces an approach, using Bayesian Networks to know the project manager's confidence on the future of its project. With this aim, a Bayesian Network is modelled using four earned value management indexes, three financial ratios, and the technical and financial capability of the contractor responsible for the execution of the project.