It might seem surprising, and a little strange, that some LIDS professors take courses in addition to teaching them. Talk to Prof. Asuman Ozdaglar, though, and the reason couldn’t be more straightforward: she adds lecture attendance to her busy schedule because she “can’t stop learning.” That passion makes Ozdaglar, one of the lab’s newer faculty members, a great fit at LIDS. Asu, as most of her peers know her, brims with a palpable energy even when sitting still. She could almost be a student with her long blonde hair and quick smile, but a minute of talking with her reveals a self-assurance that belies her age.

Asu may be relatively new to the faculty, but she isn’t new to LIDS or MIT. She came from Turkey’s Middle East Technical University to begin graduate work here, starting in RLE (Research Laboratory for Electronics), where she completed her Master’s. She then joined LIDS to work on her PhD with Professor Dimitri Bertsekas. Asu recounts that her transition from student to professor was much smoother than one might imagine. She admits that it did feel strange to walk into her first Area I faculty meeting and sit down among that small group of professors. “I wanted to sit somewhere in the corner so that nobody saw me!” She explains, laughing. “But they were very supportive.”

Asu has certainly hit the ground running during her first years on the faculty. She is eager to share her latest research, which focuses on “game theory.” This theory originated in economics, but Asu and other forward-thinking optimization theorists have taken it to engineering and, in her case, networking applications. In basic terms, her work focuses on quantifying how to get the best performance from a network, measured in terms of cost. In today’s large-scale networks, notably the internet, there is no single network planner to administer the whole thing and thus optimize its performance. For example, with a network like the internet, there are many players—service providers, customers—who have conflicting performance measures. One customer’s priority might be faster speed, another’s might be cost, while the service providers must consider both factors, as well as others like security. Asu and her students look at how to introduce control schemes to make the overall performance better for all players. This is where the game theoretic model comes in.

A typical transportation network has a great deal of traffic, and each participant—a trucking company, say—will opt for their best route, that with the minimum delay possible. It’s essentially a selfish objective for everyone involved. How will the resulting traffic pattern then look? Asu’s work can answer such a question, thereby providing valuable information in planning the overall network. What’s more, as competitors in a network conflict with one another, it might actually be in their best interest to work together and abide by common rules. Asu and her students try to determine when, and whether, such collaboration is best. This might be a strange question for those of us who drive on city streets and are accustomed to the everyday regulation of traffic lights and crosswalks. Yet when taken to a more theoretic level, it may indeed be that regulation isn’t necessary. If you let everyone pursue their own route—or best interest—will there be any real problems? Problems arising from such non-regulation are referred to as “performance loss,” which has “a catchy name,” Asu notes with a smile. It’s known as the “price of anarchy.” The term anarchy refers to the scenario in which all participants simply move on their own. Asu works to provide an overall picture of the network and measure those conflicting paths. “We suppose [there are] conflicting objectives,” she says. “If there is no conflict, there is no point, no game theoretic problem.”

Wherever such conflict exists, theorists like Asu look to determine what the optimal point of balance is, known as the Nash Equilibrium. Many laypeople may be familiar with this fundamental concept, which is named for John Nash, the mathematician profiled in the popular movie A Beautiful Mind. Asu sums it up as the point at which “nobody has any incentive to deviate. Harmony.” Yet while there are algorithms to achieve this state of equilibrium, self-interested individuals may not want to abide by them. “That’s one question I’m studying,” Asu says. She tries to form predictions by looking at the past actions of each participant, and thereby develop insight into their future moves.

This particular exploration of game theory is groundbreaking. It is highly applicable to many real-world scenarios, because a game theoretic model emerges whenever there are various self-interested users. Communication or transportation networks like those mentioned are two examples, but others include electricity markets, around which there has been much recent speculation, or wireless networks. Asu explains that game theory models are becoming sought-after in networking and computer science as important tools for resource allocation problems. Network administrators are interested for pragmatic pricing reasons, in particular. Asu’s work could ultimately help answer tricky questions for network users and providers alike. For example, what is the best way to regulate prices in network environments?

Students from various departments are taking an interest in game theory. Asu teaches a course in it for the School of Engineering, which she enjoys a great deal. She explains, “It’s one of my passions. Amazing grad students are coming to my class to see a different approach and a different set of applications than what you can find in economics departments.” She and her students work on figuring out “which parts of this theory will be useful for engineering applications.” Indeed, teaching is something Asu values deeply. She appreciates the LIDS approach to education, because “everybody cares so much about pure science,” and wants their students “to love their work.” Asu gives her teaching methods a great deal of thought. In advising graduate students, she is very involved but also expects initiative and autonomy.

Asu isn’t sure what her own future has in store. She laments the great distance from her family members and the time and effort required to visit them in Turkey. However, she remains committed to her work and her profession, and considers LIDS an ideal place to pursue both. The LIDS/MIT community is not the only to take notice of Professor Ozdaglar’s innovative work. She is a recent recipient of the prestigious NSF Career Award, of which LIDS is justifiably proud.