CRITICAL ANALYSIS OF LINEAR AND NONLINEAR PROJECT DURATION FORECASTING METHODS

The accurate estimation of the project duration at completion is still an open issue in project monitoring and control. Earned Value Management (EVM) and Earned Schedule (ES) methods provide alternative solutions, but rely on specific assumptions that may not be satisfied by all projects, thus compromising their reliability. This study provides a profile-based approach to compute the Duration Estimate at Completion (DEAC) that is general and applicable to all projects. First, two nonlinear profiles are introduced, and their DEAC formulae are derived. Then, the ES, the two profile-based, and the EVM forecasting methods are tested on a portfolio of 56 engineering projects by computing each method’s DEAC at 5%-progress stage. The collected results are analyzed under both the portfolio and individual project levels. Findings show that the extra effort in using the profile-based methods is valuable in improving the DEAC forecasting performance for specific projects. The correlations between the four DEAC methods’ accuracy, the project cumulative work profile, and the final schedule delay are analyzed further to investigate the drivers behind the four approaches’ performance. The results confirm that one DEAC method should be prioritized over the others according to the project S-curve shape.