Icon Legend

This session is not in your schedule.

This session is in your schedule. Click again to remove it.

116 Views

Abstract for presentation (Poster or Podium) with a Paper in the Conference Proceedings

Highway Pavements

Forecasting Highway Maintenance Cost at Early-Stage using Machine Learning

Monday, June 17, 2024

10:15 AM - 11:30 AM ET

Room: Galleria

Primary Author(s)

YP

Yejee Paik

Ph.D. Student
Georgia Institute of Technology
Alpharetta, GA, United States

Co-Author(s)

FC

Frederick Chung, n/a

Ph.D. Student
Georgia Institute of Technology
Atlanta, Georgia, United States
BA

Baabak Ashuri

Associate Professor
Georgia Institute of Technology
Atlanta, Georgia, United States

Submitter(s)

YP

Yejee Paik

Georgia Institute of Technology
Alpharetta, Georgia, United States

Abstract: Accurate cost estimation has been a notable challenge for the development of cost-effective highway maintenance plans. While multiple variables, including market indices, have been used to attain higher accuracy of cost estimation models, the difficulty escalates for long-term forecasting. Especially in the prediction of early-stage of projects, the limited amount of information and the unpredictability of market conditions make it intricate to achieve the desired accuracy. This paper applies machine learning (ML) to develop a highway maintenance cost forecasting model using only the early-stage project details. Georgia Department of Transportation (GDOT) p</span>avement data from 2017 to 2021, consisting of ‘project-related’ and ‘pavement-related’ variables, is input to ML methods such as linear regression, random forest, extra trees, support vector regression, and other ensemble methods. Extra trees regression showed the lowest mean absolute percentage error of 19.16% after 5-fold cross validation. This study will work as a reference for DOTs to implement ML in the long-term maintenance cost prediction, and ultimately be utilized in the tradeoff analysis of determining the optimal management strategies.

Learning Objectives:

Attendees can expect to learn the following from this session:

recognize the difficulty of long-term construction cost forecasting with early-stage projects.
classify the two categories of variables which can be input to the highway maintenance cost estimation model regarding early-stage projects.
employ machine learning methods for highway maintenance cost forecasting.
identify machine learning algorithm with best performance in terms of mean absolute percentage error in given conditions.