Dartmouth Engineering Students Take First Place in Prestigious Operations Research Competition

PhD student Prabhat Hegde and BE student Alan Ngouenet '25 accept the award at the MOPTA conference. PhD student Yueyun Xia pictured right. (Photos courtesy of Lehigh University and Yueyun Xia)

"I'm super grateful for my brilliant teammates who made the late nights working on this competition energizing and fun," said BE student Alan Ngouenet '25 about his PhD teammates, Prabhat Hegde and Yueyun Xia. "It was such a fun and exciting experience, and it couldn't be achieved without my teammates," said Xia. 

The trio took on a real-world challenge from Transavia Aircrew Scheduling which is working toward transforming their fleet of aircraft from Boeing to Airbus. "This transformation process has provided a series of complex challenges," stated Transavia. "One of our more intricate puzzles is our cockpit crew training," and they requested a solution that would "ensure a seamless training schedule that maintains operations readiness … with the minimum cost to our company."

"It’s a truly proud moment for Thayer and for our research group," said Professor Vikrant Vaze, a principal investigator in Dartmouth's Operations Research (OR) Group, after the winner was announced.

The 2025 competition attracted approximately 20 OR teams worldwide. Three finalists were then invited to the annual MOPTA conference to present their solution. We talked with winning team member Hegde about the experience:

How did you get involved in this competition?

I've always wanted to take part in one, but haven't had the bandwidth. I saw this on some of the operations research forums and posted it on our internal lab Slack group. Yueyun's in her first year, she's still doing classes, and she told me she was interested in taking part as a course project. It just so happened that there was an overlap in the material of the class and what we were trying to do.

Why were you so interested?

What was interesting was how differently people modeled the same problem—that's super fascinating to me. Key decisions were, how many aircraft do you ground and when do you ground them? And also, who you train and what type of training? The solution approaches were all different and I think this is the beauty of abstracting reality into models. You cannot capture the complexities of reality. You can only approximate it in a model. For any big problem, wise people say, "All models are wrong, some models are useful." It's just because some models capture the abstraction better than others.

Were you surprised when your team won?

I was fairly confident. I'm a sixth-year PhD and pretty well-versed in the field, and I know a fair bit about these types of models. We wrote out the model from scratch, spent a lot of time on it, and the day we were submitting the report, all three of us were in my office until midnight editing, polishing, and tying it all together.

What gave your team the edge?

I think we won primarily for one reason—we understood what was important to the company. When they had the info sessions, they told us a few specifics about the problem, that there's the possibility of hiring external pilots versus grounding a certain number of aircraft. Then they said they really, really, really do not want to hire external pilots, and they gave us a cost matrix for grounding only up to ten aircraft. There was no dollar amount for the cost of grounding more than that. So, we spent a long time trying to ensure that that constraint was met. I'm not sure what else was the clincher, but based on the final presentations, I felt like we really responded to the stakeholder needs more than the theoretical aspect of it, which I think really helped us.

Also, there was something in the way we formulated the model that made our solution very quick—our model solved on a local computer within 0.1 second and was a good solution. But again, in each solution there were interesting things. I felt like the team that was third place explored a lot more theoretically. The team that placed second spent a lot of time not on the model, but on creating a visual interface where they made up thousands of fake pilots, assigned them names, languages they speak, where they go—completely cooked up data for their interactive app.

What was your favorite part of the competition?

For me, personally, it was a great mentorship experience, and it felt nice to have a good team effort. All three of us brought something to the table. Alan spent a lot of time on the visual app interface. I just gave him some instructions and a rough outline of what I'd like to see and he went and did it. Yueyun was very, very helpful with the model formulation. We spent quite a bit of back-and-forth correcting things. Some of the things I did not catch, she was good at catching and correcting, and was instrumental in keeping that ten aircraft cap. Yueyun couldn't make it to the final, but after we got first place, Alan came and gave me a big hug and said, "I couldn't do it without you." It was big in that way.

What do you want people to know about OR?

Pretty much everything you do in your life is governed by OR. From your Amazon package delivery to your electricity—what that electricity is priced at, how much is supplied, what power goes on at what time—underneath all of that is someone writing a linear program or an integer program and basically governing your life. And take any mode of transport—from the price of your ticket to what seat you're assigned, to the crew that's on your airplane—all of that's governed by different models interacting with each other at the end of the day.

What's next for you?

I'm waiting to set a thesis defense date before I start looking for jobs. I probably would love working at an electric grid operator, something along those lines. I really like working on electricity markets, pricing, and planning the electric grid of the future.