Final Presentations from this year's Capstone Projects
We wrapped up another Client Project Challenge (CPC) class this week. This is our name for our Industrial Engineering capstone class. At the suggestion of our advisory board, we do this in the spring quarter (~end of March until early June) of junior year. By doing it in their junior year, the students get a real-world experience before their internships and job searches.
We had thirteen projects this spring. The final presentations and Q&A were great. The class is set up so that the students run the projects. I last see them (and give advice) with about 3-4 weeks to go. Many projects (like the ones we’ve all been on) usually have a lot to do during those last 3-4 weeks.
The students did a great job! Here is a quote from a faculty member who saw the projects for the first time during these presentations: “…those were an exceptional group of projects with great work that was presented well.”
I couldn’t agree more with that quote. I also think the student teams learned a lot. Here are some of the highlights for me from the final presentations:
Explaining the results. A few projects required detailed explanations of why the algorithm made the suggestions it did.
Building software solutions. Several projects had to deliver a software solution. IE work is more than just the algorithms. Streamlit seemed to be a popular solution.
Stitching together IE algorithms. Many projects required multiple algorithms to answer the business question. For example, we had a project that required demand forecasting before optimizing the labor scheduling, or needed inventory analysis to help with transportation mode selection, or one that required cutting stock analysis to help with inventory.
Simple can be good. Sometimes, building sophisticated models helps you find simple solutions that add value. This is a good lesson. As engineers, we often don’t feel we’ve done our work unless the recommendation involves a complex model.
Working with incomplete data. This is a great lesson that causes pain during a project, but we all deal with it when working on real-world problems.
Using BERT. Two projects used some version of BERT, a family of state-of-the-art deep learning tools pre-trained to understand language in context. I found it funny that the students thought of BERT as just another tool.
Savings and benefits come in many forms. We tend to think about savings in just dollars. This year’s class had a mix of savings from improved operational efficiency, improved safety, automating tedious work, restructuring workflows, rearranging work shifts, insights about the business, and, of course, a few with some concrete dollar savings.
In the coming days and weeks, I hope to get client-specific stories and share the ones we can.
Congratulations, Dr. Watson. You make IE sound fun! May the students go on to bigger and greater things!