Leading Thoughts: Where Innovation Happens

Dean Alexis Abramson talks with Professor Vikrant Vaze. (Photo by Rob Strong '04)

How do you define operations research?

VAZE: I define it as any method that helps you solve a decision-making problem using mathematics and computing. That can involve optimization, simulation, predictive modeling such as machine learning. It can involve statistics, game theory, mechanism design—all of these are examples of methods you can use.

Everything that we do comes down to three steps. First, there's the step of understanding the practical domain—this might involve the physics or the economics or the human behavior side of the problem. Once we understand that, we go into the question of how to model this practical system. The model will involve a bunch of equations and inequalities at the end of the day, but there is a lot of skill in getting there. It's an art as well as a science. That's where much of the innovation might happen. And then the third piece is how you solve these complicated mathematical models in a way that's optimal. That's again where innovation happens.

Typically, we end up doing a few iterations of this whole process. Many times you might talk to people, learn about the practical context, try to do a math model, and then realize, 'No, we have to change that.' That sort of feedback can happen at all these steps. But this is how all the problems go: the practical context, the math, the algorithms and computing, and then looping it back to the decision-makers.

I travel a lot. Can you walk me through how operations research is being applied to me, the flyer?

VAZE: We've done a lot of work where we asked ourselves and the airlines: 'How can we make the traveler experience better?' This can lead to a win-win because if travelers are better off, then they're willing to pay more. Most of the customer-centric work has been done in the context of pricing and revenue management, but until recently people didn't zero in on the customer behavior aspects of the scheduling process. Schedules are often designed to meet certain criteria that indirectly consider buying behavior, but they don't directly ask the real question of would the customers buy this? That is a big departure we've undertaken over the last decade. Now when I try to understand how good a schedule is, I'm going to include models from behavioral economics that take into account customer choice.

How have you expanded your research into other applications?

VAZE: One example, with the support of Irving Institute for Energy and Society as well as the Dickey Center for International Understanding, is a project called Powering Peace. This is about areas in the world ravaged by continuous conflict and a very serious problem associated with how we can bring more peace. Central to that challenge is what drives the energy dynamics, in particular, where many are focused on a diesel reliant economy. Diesel supply chains are very messy, complicated, and prone to attacks. It's a very different context because most people in these regions have no source of energy available. So, this project could drive access to electricity.

And risk on investment takes on a whole different meaning in these fragile, war-ravaged places. With these nuances in mind, we have multiple objectives and stakeholders, including several different United Nations departments involved in primarily peacekeeping operations. There is the local industry, the local governments of these states and regions, and investors. There are people interested in investing in these places, but they want to have a better analysis of the chances of these investments working out. What we are trying to do is build a big data-driven, decision-making framework where we provide tools to optimize their decisions. Then, we have tools that will take in those decisions, optimize the overall design of the system, and evaluate and present tradeoffs to stakeholders.

How do you involve students in your work?

VAZE: We have honors thesis students, MEM students, and our six PhD students working side by side. We have several [First Year Research in Engineering Experience] FYREE students, and then we have maybe two or three others who heard about the program and asked to be involved. Our freshmen are working on three different problems spread out across transport, logistics, healthcare systems, and energy and climate. When you put all these individuals with such an amazing variety of lived experiences in the same room and we start talking, something magical happens: Completely new ideas come out. That's one of the most exciting parts of my day.