Georgia Institute of Technology
Data-informed Change Order Risk Management in Transportation Infrastructure Using AI-Enabled Insights
Abstract
dc:description.abstractThis dissertation investigates how persistent change orders in transportation infrastructure projects can be more systematically understood and managed through data-driven analytical approaches. Using project-level data from a state department of transportation in the United States, it integrates statistical, survival, and artificial intelligence-based methods to analyze change order risks and their impacts on project performance and to develop a novel impact profiling approach for more risk-informed project management. The first study empirically assesses variability in project performance and change order impacts across different project delivery methods. Findings show that design-build projects generally achieved better project effectiveness and lower change order frequency than traditional design-bid-build projects, while change order risk patterns and impact characteristics varied by delivery method and reason type. The second study explores the recurrence dynamics of change orders throughout project phases using recurrent event modeling. Results indicate that larger projects, the design-bid-build delivery method, higher contingency levels, minor contract duration extensions resulting from prior change orders, and Fall-season change orders were associated with elevated recurrence risks. The accumulated recurrence of change orders also contributed to greater overall schedule and cost impacts toward project completion. The third study introduces an explainable artificial intelligence (XAI)-based impact profiling framework to predict and interpret the schedule and cost impacts of change orders. Tree-based ensemble learning models demonstrated strong predictive performance, while XAI-based interpretation identified critical factors and value-dependent impact patterns associated with change order-related disruptions. The proposed framework further provides an interpretable visualization of change order risks by jointly profiling schedule and cost impacts, improving the interpretation of factor-specific disruption patterns. This dissertation contributes to the body of knowledge in transportation infrastructure project management by advancing data-driven approaches for understanding and managing change orders. Collectively, the findings offer practical and interpretable insights into change order-related risks, supporting proactive risk mitigation and more resilient project delivery.
Degree
thesis:*- Name thesis:degree_name
- Building Construction, PhD
- Department thesis:degree_department
- Building Construction
- Grantor
- Georgia Institute of Technology
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Oh, Jeongyoon
- Advisor dc:contributor.advisor
-
- Ashuri, Baabak
- Committee members dc:contributor.committeemember
-
- Messner, John
- Yang, Eunhwa
- Wen, Jing
- Yan, Karen
- ORCID person:identifier.orcid
Rights
- Language dc:language.iso
- English
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1853/82175