Description

The State Highway Operation and Protection Program (SHOPP) is crucial for maintaining California’s 15,000-mile state highway system, which includes projects like pavement rehabilitation, bridge repair, safety enhancements, and traffic management systems. Administered by Caltrans, SHOPP aims to preserve highway efficiency and safety, supporting economic growth and public safety. This research aimed to develop robust cost-estimating models to improve budgeting and financial planning, aiding Caltrans, the California Transportation Commission (CTC), and the Legislature. The research team collected and refined comprehensive data from Caltrans project expenditures from 1983 to 2021, ensuring a high-quality dataset. Subject matter experts validated the data, enhancing its reliability. Two models were developed: a statistical model using exponential regression to account for non-linear cost growth, and an AI model employing neural networks to handle complex relationships in the data. Model performance was evaluated based on accuracy and reliability through repeated testing and validation. Key findings indicated that the new models significantly improved the precision of cost forecasts, reducing the variance between predicted and actual project costs. This advancement minimizes budget overruns and enhances resource allocation efficiency. Additionally, leveraging historical data with current market trends refined the models’ predictive power, boosting stakeholder confidence in project budgeting and financial planning. The study’s innovative approach, integrating machine learning and big data analytics, transforms traditional estimation practices and serves as a reference for other state highway programs. Continuous improvement and broader application of these models are recommended to further enhance cost estimation accuracy and support informed decision-making in transportation infrastructure management.

Publication Date

11-2024

Publication Type

Report

Topic

Transportation Finance

Digital Object Identifier

10.31979/mti.2024.2365

MTI Project

2365

Keywords

Cost estimating, Neural networks, Regression analysis, Project portfolio management, State Highway Operation and Protection Program (SHOPP)

Disciplines

Finance and Financial Management | Technology and Innovation | Transportation

Download

Additional files available below

Share

COinS