Doctoral Dissertations
Date of Award
5-2025
Degree Type
Dissertation
Degree Name
Doctor of Philosophy
Major
Industrial Engineering
Major Professor
Mingzhou Jin
Committee Members
Mingzhou Jin, Hugh Medal, Xueping Li, Hyun Kim
Abstract
This dissertation develops a comprehensive framework to study freight intermodal transportation, focusing on a holistic understanding of the interactions between trucks and railroads. The research is structured into three main chapters, dedicated to the following topics: life-cycle benefit-cost analysis, two-stage stochastic optimization of location-allocation problems, and stochastic optimization of container movement problems.
The first chapter conducts a comprehensive Life-Cycle Benefit-Cost Analysis (LBCA) comparison of highways and railroads to evaluate the nationwide impacts on economic, social, and environmental impacts across the life cycles of transport infrastructure and equipment. The study examines both actual and maximum capacity flows to identify cost-effective, sustainable investment options. This tool can support stakeholder-specific decision-making, depending on whether their goals prioritize specific impacts or broader impacts.
Chapter Two explores an important intermodal system case study of inbound freight in the U.S. Southwest region, which originates from California's San Pedro Bay port complex (SPPC). To improve system efficiency, we propose a strategy of expanding inland logistic centers for container classification across four states—California, Nevada, Arizona, and Utah. A two-stage mixed-integer stochastic programming model is developed to manage demand uncertainty, and sensitivity analyses are conducted to evaluate performance under varying cost and demand scenarios.
In Chapter Three, we extend the scope of the preceding model to capture the operational dynamics of inbound freight container movement within the SPPC intermodal system. This expanded model incorporates key factors such as container flow, mode selection, operational timing, and lead time considerations. The problem is formulated as a Markov Decision Process and solved using deep reinforcement learning techniques, including Deep Q-Learning and Advantage Actor-Critic. The goal is to minimize container dwell time at ports and warehouses, thereby increasing system efficiency, reducing transportation time and costs, and enhancing the overall resilience and adaptability of the intermodal logistics network. The trained models demonstrate the ability to learn effective container movement strategies, providing valuable decision support for planners seeking to optimize operations, reduce delays, and respond to fluctuating conditions.
Recommended Citation
Rattanakunuprakarn, Sarita, "Intermodal Freight Transportation Management: Life Cycle Benefit-Cost Analysis, Stochastic Facility Planning, and Reinforcement Learning for Container Movement. " PhD diss., University of Tennessee, 2025.
https://trace.tennessee.edu/utk_graddiss/12414