Author: Matthew Walters Major: Industrial Engineering
During my sophomore year in college, I knew that I wanted to conduct undergraduate research, but I didn’t know how to find a professor to work with. Luckily, the Industrial Engineering department offered an Honors Research Experience class for underclassmen who wanted to pursue research. Through that class, I was able to find a topic I was interested in, and I was able to meet Dr. Chase Rainwater who has been an awesome professor to work with. My research seeks to understand how collaborative logistics can reduce the cost, miles traveled, and carbon emissions of trucking. Collaborative logistics is where competitors work together to serve a common customer. This includes the sharing of resources and information to reduce the operating costs of all parties involved.
Being only a sophomore when I first chose this research topic, I still needed to take a class called operations research in order to be able to have the necessary skills to start the technical part of my research. So I began literature review to understand all the different types of companies in a supply chain, and how they interact with each other. I read a book called “The Physical Internet” by Eric Ballot, Benoit Montreuil, and Russell D. Meller. This book lays the groundwork for how the future of logistics should operate. It speaks of standardized containers, optimized routing protocols, efficient handling facilities, and much more. After my literature review and completing the operations research course, I now had the necessary knowledge to begin the technical part of my research.
I began by creating a mixed integer linear programming model (MILP) that models how a set of manufacturers, customers, and third party logistic companies would interact within a supply chain. This model will optimize the decisions that these companies must make in a given time period such as how to route an order, how many trucks to use, and which distribution center to use. I created two models: a collaborative and a non-collaborative model. The main difference is that the collaborative model allows for all companies to share their trucks and distribution centers.
After the two models were created, I then created a Python program that will randomize the supply chain data. The program will take in multiple inputs such as the number of time periods, manufacturers, retailers, and third party logistics companies. The program will then calculate and determine data such as facility location, demand, production, number of trucks, and more. This is the data that will be used as the parameters for the two models that I created.
The next step in my research will be the experimental phase. I will first use the two collaborative and non-collaborative models that I have created to solve the decisions of the supply chain that I have randomly generated. I will then compare the total cost and miles driven between the two solutions produced from the two models. I will also adjust the parameters of the supply chain such as the number of companies, the size of the companies, and the number of trucks and distribution centers that companies own. This experimental phase will help discover the type of supply chains that can most benefit from collaborative logistics.
After I complete the experiments, I will conclude my undergraduate research by defending my thesis. I am excited to be able share what my research will discover. It has been great working with Dr. Rainwater throughout these past two years, and I am excited to continue conducting research with him next year as I return for graduate school.