Investigation of the Earned Value Method (EVM) Applicability for Construction Operations Affected by the Learning Phenomenon

Abstract

The Earned Value Method (EVM) has been extensively applied for the analysis of construction projects. However, in cases where the productivity is not constant, but rather varies due to accelerations attributed to the learning phenomenon, it is challenging to assess the implications on productivity estimation and forecasting. In that sense, such an investigation is crucial for the scheduling and coordination of the remaining works in a realistic manner. The purpose of this paper is to compare the progress reporting results using the Earned Value method both for the theoretical project time schedule (without learning) and the actual on-site scheduling following the learning curve. A real, large-scale infrastructure project is used as a case study. The research method involved the “transformation” of productivity data to cost data, with the purpose of quantifying the productivity improvements through the use of statistical learning models. The straight-line model was used, due to its wide acceptance in related studies. An algorithmic approach is developed and assessed via, inter alia, the Schedule Variance (SV) and the Cost Variance (CV) indices. The results of the research indicate that EVM is significantly affected by the learning phenomenon, which, if neglected, leads to ineffective decision making procedures, regarding the deployment of project resources.