Forecasting Final Cost and Budget of Construction Projects

Abstract

This paper describes a computational approach to calculating the final cost and budget of a construction project, which is based on the data normally developed by a cost system during the life of the project. While most cost systems are able to collect total budget and to-date cost, budget, and quantity data, there is often less capability to use this data to forecast the final cost of a project. In many cases, the cost forecast for unstarted work does not reflect the overruns or underruns that are recorded early in the life of a project. This paper presents an algorithmic approach that has been tested using the data from 121 completed construction projects, which calculates reasonably good cost and budget forecasts over a wide range of project types and changed conditions. The proposed method is compared to two other methods and is found to be superior in accuracy, timing, and stability. The forecasting method, sliding moving average, is general in its approach and may be useful for other situations in which predictions of limited duration time-series data are desired.