Combination of Growth Model and Earned Schedule to Forecast Project Cost at Completion

Abstract

To improve the accuracy of early forecasting the final cost at completion of an ongoing construction project, a new regression-based nonlinear cost estimate at completion (CEAC) methodology is proposed that integrates a growth model with earned schedule (ES) concepts. The methodology provides CEAC computations for project early-stage and middle-stage completion. To this end, this paper establishes three primary objectives, as follows: (1) develop a new formula based on integration of the ES method and four candidate growth models (logistic, Gompertz, Bass, and Weibull), (2) validate the new methodology through its application to nine past projects, and (3) select the equation with the best-performing growth model through testing their statistical validity and comparing the accuracy of their CEAC estimates. Based on statistical validity analysis of the four growth models and comparison of CEAC errors, the CEAC formula based on the Gompertz model is better-fitting and generates more accurate final-cost estimates than those computed by using the other three models and the index-based method. The proposed methodology is a theoretical contribution towards the combination of earned-value metrics with regression-based studies. It also brings practical implications associated with usage of a viable and accurate forecasting technique that considers the schedule impact as a determinant factor of cost behavior.