Enabling Proactive Waterway Transportation

Watching the simulation unfold in real-time.

Author: Ashwin Narayan | Major: Industrial Engineering | Semester: Spring 2023

Key Information: I am Ashwin Narayan, a senior Industrial Engineering student in the College of Engineering, and my mentor is Dr. Haitao Liao, also of the Industrial Engineering department. My semesters of research are Spring and Fall of 2023, and after graduating in May 2024, I hope to apply my Industrial Engineering skills in industry.

Blog:

Inland waterway travel accounts for a significant portion of the transportation of goods for many industries, including agriculture, the chemical, and the natural resource industry. An effective simulation tool modeling components of a waterway network could be utilized both by the private and public sectors. Companies could use this tool to optimize travel paths for barges containing consumer products to ensure that those products can be delivered on time. The state government could utilize this tool to determine where, when, and why waterway infrastructure (i.e., locks and dams) needs maintenance. Both of those applications would require a simulation tool which accurately depicts the complexity of the waterway network, and which is capable of using historical trends and statistical distribution analysis to perform predictive analysis on the likelihood of infrastructural failure.

I chose this topic because statistical modeling is at the heart of Industrial Engineering, and I wanted to apply those techniques to a tangible real-world problem. Studying simulation models and performing statistical analysis would enhance my statistical modeling skills while providing a benefit to users in both government and industry. My mentor, Dr. Haitao Liao, taught my first Engineering Statistics course at the university, and since I thoroughly enjoyed his course, I wanted to learn more about his ongoing research. In one of my research courses, he gave a presentation on statistical modeling of the navigational inland waterway network in this region, and I knew this was something I would enjoy learning more about.

The primary challenge for me as I learned more about this topic was understanding the statistical complexity behind the simulation model. Both of my partners on this project were PhD students who had taken many more statistics and simulation courses than I had. Familiarizing myself with topics at a graduate level as an undergraduate student was very difficult. Fortunately, both of my partners were understanding of my situation and allowed me ample time to conduct background research without feeling that I had make a contribution immediately. Another difficulty that I faced was building reliable models with incomplete data. To understand trends and why they occur, the data must be complete. In the case of our project, and likely every other project, the data we studied was never complete, and to overcome this, we had to make a few assumptions. While this isn’t ideal, if the assumptions are intelligent and the researchers clearly document them, they won’t have too much of a negative impact on the analysis. On a personal level, a major challenge was creating a schedule and appropriately pacing myself throughout the project. Balancing expectations for research deliverables became hectic during the busier parts of my semester, so being upfront with stakeholders about what I could accomplish was critical. Flexibility, or balancing my research workload around my coursework and extracurricular schedule, was critical, and throughout the course of the semester, I became better at allocating my time towards my research. Finally, I really enjoyed applying techniques that I had studied in class to my research project. This project gave me the opportunity to utilize statistics and simulation techniques which I had primarily learned about in a theoretical context. I found that applying topics I learned in class to real-world problems was very fulfilling.

My faculty mentor was active in conducting meetings with my graduate student partners and myself. These meetings allowed us to describe updates we had made and receive questions from other faculty members knowledgeable on the topic. He was always available for advice and consultation whenever we had questions. My primary partners were two graduate students who were also working on the simulation model. I benefited greatly from their technical expertise and frequently met with them to discuss questions, issues with the project, and data analysis. Since there was a large technical gap between myself and my partners, I heavily relied on them to teach me about how to operate the simulation model and perform comprehensive analysis of the results. Thankfully, both of them were eager to help.

I didn’t travel for project purposes this past semester but am interested in traveling to an Industrial Engineering conference to share some of my discoveries. Additionally, I would like to travel to a lock site in Arkansas to view a portion of our inland waterway infrastructure first-hand. In the spring, I hope that this simulation can be handed off to stakeholders who can use this to make proactive decisions when routing their vessels along the waterways. While my plan after graduating is to work in industry, I would like to revisit this my research in a few years to understand how it is being used and how it has been modified to reflect changes that may occur between now and then.