A Systematic Literature Review on Cost Forecasting Techniques for Improving Estimates Accuracy in Construction Projects

Abstract

In construction, one measure of accuracy is the deviation obtained between estimate at completion and the most responsive tender figure. While different forms of forecasting techniques have been used, the era of artificial intelligence has also availed the industry the tools to improve construction cost estimation. However only a few studies address accuracy within the defined context of accuracy. The goal of this systematic literature review is to map out knowledge territory, highlight means of improving accuracy within the cost forecasting domain. From a total of 133 articles retrieved from EBSCOhost, Google Scholar, Scopus databases and ASCE library, 93 articles published between 1976 and 2022 were considered for the study. The review reports the contributions of extant literature which have so far focused on construction forecasting tools and techniques deployed to improve accuracy. A key highlight of the findings is that the reference to accuracy is given to models’ ability to manipulate data (independent variables) and not in any sense comparing a final figure to one forecasted. This gap highlighted is the goal of an ongoing study aimed at examining current efforts at improving accuracy and thus developing an effective cost forecasting model using improved techniques. Through the analysis extant literature on cost forecasting, this review concludes that a robust big data analytics approach is needed to manage the shortcoming of existing techniques while taking into cognizance the series of events that are correlated and influence cost across the project life cycle. The review ends with a recommendation to consider a shift away from over-reliance on model development to training and operationalising models to adequately capture factors that influence cost overrun.