A residual Grey prediction model for predicting S-curves in projects

Abstract

S-curves are usually taken as expression of project progress and have become a requisite tool for project managers through the execution phase. The common methodology for predicting S-curve forecasting models is based on classifying projects into groups and producing a standard S-curve for each group using multiple linear regression techniques. Traditional regression models taken to fit individual projects require a large amount of data and make many strict assumptions regarding statistical distribution of the data. The grey system theory, however, is well suited to study the behavior of a system with incomplete information or limited amount of discrete data. Easy of use and accuracy, two significant criteria for project managers when choosing a forecasting model, are considered two additional attributes of the grey system theory. This paper proposes a residual Grey prediction model to forecast the actual cost and the cost at completion of a project based on the grey system theory. Results show that the accuracy of the forecasting model is highly efficient.