A Comparison of Exponential Smoothing and Time Series Models for Forecasting Contract Estimates-at-Completion.
Richard J. D'Accardi · 1996
Abstract : Several models are embedded in Performance Analyzer (PA) that provide reasonable forecasts of program cost using information contained in contractor-provided monthly Cost Performance Reports (CPR). Considering the nature of this data, both Exponential Smoothing and Time Series models are good candidates to forecast Estimates-at-Completion (EAC). The thrust of this report is twofold. First, to expand the currently available forecasting models that are components of PA to include Time Series and Exponential Smoothing models. Second, for known contract cost performance data, using PA generated forecasts as a benchmark, a comparison of the models is made to highlight the differences between them and to provide a basis for choosing the best model for determining EAC's.