Improving the way that WSDOT performs business is an important objective to pursue. The objectives of this research were to develop tools that will monitor the contractor’s performance during construction in order to detect any unsatisfactory progress, and to develop tools that will improve the time and cost prediction of highway projects in order to reduce time and cost overruns.
To achieve the first objective, the research surveyed other state DOTs about how they measure and evaluate work progress and contractor performance. The survey showed that a formal progress measurement and performance evaluation process is lacking in many states, and that there is an apparent lack of progress charts for measuring contractor performance.
By using WSDOT historical project data on actual payment estimates and the elapsed working days of each estimate in each project, the current research developed minimum performance bounds and average performance bounds for a set of successfully completed projects using regression analysis. Performance bounds were developed for all projects and for clusters of projects grouped in categories based on quantities of asphalt concrete pavement/hot mix asphalt (ACP/HMA), contract value, project duration, and project miles.
Time and cost prediction models were developed through the application of general multiple regression analysis, ridge regression analysis, and nonlinear partial least-square regression analysis on WSDOT historical project data. The models were developed on the basis of a number of major variables in pavement projects, including project duration (working days), final contract value (paid-to-contractor dollars), ACP/HMA quantity (tons), grading (tons, cy), surfacing (ton), and the number of project highway miles.