Combined machine-learning and EDM to monitor and predict a complex project with a GERT-type network: A multi-point perspective

Abstract

Project monitoring is an important topic in the field of project management. This paper proposed a hybrid model of stochastic EDM (Earned Duration Management) and machine learning for a complex project with a GERT-type (Graphical Evaluation Review Technique) network. First, an EDM analysis for a GERT-type network is conducted based on the Monte Carlo simulation. Then, considering the synergistic effect of multiple control points, a new project monitoring algorithm is designed to identify deviations and generate early warning signals. Consequently, several regression algorithms are utilized to estimate the expected project completion time. Besides, the effect of different control points, different project due dates and machine learning algorithms on model performance is explored. An illustrative case study is used to demonstrate the effectiveness of the proposed model.