Poster 20-17

ENHANCING THE ACCURACY OF CONSTRUCTION COST ESTIMATES FOR MAJOR LUMP SUM (LS) PAY ITEMS AND GENERATING A MORE-ACCURATE LIST OF PAY ITEMS THROUGHOUT THE DESIGN DEVELOPMENT PROCESS


PI: Baabak Ashuri

Co-PI(s): Minsoo Baek

Institution(s): Georgia Institute of Technology


Abstract

State departments of transportation (DOTs) encounter a critical challenge in estimating accurate cost estimates for major lump sum (LS) pay items, such as Traffic Control and Grading Complete, due to incomplete project information during the early stages of project development. To estimate prices for LS pay items, cost estimators and designers in state DOTs apply engineering judgment using knowledge from similar projects from the past and reach out to subject matter experts for providing additional resources. However, researching similar projects for finding appropriate estimates for the LS pay item is not a simple endeavor and leads to significant inaccuracy of cost estimates. A need exists to develop new methods that are capable of capturing key information from project documents and incorporating the complex nonlinear relationships between input and output variables in developing prediction models for LS pay item prices for highway projects. Thus, the overarching objective of this research is to develop forecasting models to estimate the prices of the Traffic Control and Grading Complete LS pay items using advanced text mining and machine learning algorithms that detect key patterns of information generated during project development and provide higher accuracy in cost estimates. In this research, a forecasting model for the prices of the Traffic Control and Grading Complete LS pay items was developed using machine learning algorithms (i.e., random forest, bagging, k-nearest neighbors, and stacking regressor). Furthermore, a web-based application tool was developed in a Python environment to provide project designers developing cost estimates with a data-driven tool for estimating the prices of the Traffic Control and Grading Complete LS pay items.


Please comment below with any statements or questions you may have. Also let GTI if you would be interested webinars or presentations on similar topics.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer