PREDICTING PRODUCTION RATES IN THE WISCONSIN HIGHWAY CONSTRUCTION INDUSTRY USING COMPUTER TOOLS
Abstract
Poor estimates of project contract completion time can produce dire consequences. Awarding too much time to complete a project creates conditions where contractors may be encouraged to perform work less aggressively, while awarding too little time can increase prices and decrease the amount of bids received. Traditionally, the Wisconsin Department of Transportation’s (WisDOT) methods for determining contract completion time have relied upon production rates published in their decades old Facilities Design Manual (FDM). However, the FDM has not been updated to reflect advancements in technology, and the use of its production rates has fallen out of favor with WisDOT staff. This study improved upon the prototype version of a productivity estimation tool (PET) created to provide WisDOT estimators with more reliable productivity information.
WisDOT’s PET uses models created using stepwise linear regression to identify significant project specific factors in order to estimate productivity on future projects. The prototype of WisDOT’s PET suffered from limited data and omitted several miscellaneous highway construction activities. This study improved upon the PET by gathering additional data, including 26 new miscellaneous activities, enhancing the visual means of displaying productivity information, identifying additional factors for estimators to consider when determining contract time, and creating a web domain to gather productivity data from WisDOT field engineers in the future. Applying the PET and incorporating new data into its models as projects are completed will allow WisDOT estimates of contract time to improve over time.
Subject
regression
estimation
prediction
productivity