An Investigation and Analysis of Non-Linear Project Cost Forecasting in Earned Value Analysis Using the Beta Family of Equations

Abstract

As part of the subject area of cost-schedule control, a non-linear method for forecasting project costs using the beta family of distributions is investigated. Linear methods have frequently been characterized as unrealistic. Cumulative and distribution curve formulas for cost parameters in EVA are derived with the beta family of equations using elementary calculus. The technique for using the beta family of distributions is illustrated with EVA rigor in mathematical terms. An example project is presented throughout for making forecast calculations. Some limitations exist for the application of the equations, thereby imposing constraints on the type of projects and project conditions that can be analyzed. The equations can be used at each update for interim forecasts and not for end-of-project forecasts. For the examples considered, non-linear forecasts can vary from the linear by 16.14%, depending upon the particular data characteristics.