I was recently in the wonderful city of Istanbul, the venue of the Fourth World Congress of the Game Theory Society. Just like the Olympics, this is an event held every four years and is, in fact, sometimes referred to as the Game Theory Olympics — even though there are no medals to be awarded.
The tagline for the conference, hosted by the Bilgi University late last month, was “Istanbul Welcomes Beautiful Minds”. That was an obvious allusion to the four Nobel Prize winners — Eric Maskin, Roger Myerson, Reinhard Selten and, of course, the original Beautiful Mind, John Forbes Nash, Jr.
They were all there.
Attending a full day of a meet like this can be mind-numbing. After four days of four plenary sessions, four semi-plenary sessions and 498 invited papers — well, you’ll certainly start getting a hang of what’s happening at the frontiers of game theory research today.
The conference covered all the usual suspects like evolutionary games, mechanism design and contract theory, social choice, voting, and political economy. But two emerging areas that seemed to receive special emphasis were algorithmic game theory and networks.
New Games
Algorithmic game theory, as some researchers in the field put it, is the confluence of two ideas put forth by the great Hungarian-American mathematician, John von Neumann: Game theory and Algorithms.
Game theory basically looks at how human beings behave in situations where the final outcome depends on the actions of more than one person, as in a game of chess or tic-tac-toe. An algorithm is a finite and precise set of instructions to achieve a certain desired outcome. In algorithmic game theory, the objective, then, is to design algorithms for particular strategic situations, based on predictions of human behaviour provided by game theory, and to achieve the desired outcomes algorithmically.
The beauty of this approach lies in the fact that algorithms behave like the true homo economicus — the perfectly rational agents whose existence is usually assumed by economic theory.
There were a number of papers at the conference, focusing on the complexity of obtaining computational solutions to game-theoretic problems in domains such as logistics. Besides, we had the normal slew of papers on auctions and mechanism design — major areas of interest to game theorists and computer scientists alike.
The study of networks and its applications in economics is a little over a decade old. In contrast to the research on this subject in sociology and physics, the first question economists seek to ask is how networks are formed in the first place. What is the role of incentives in their formation and do networks that are individually stable coincide with the ones that are socially optimal?
While some papers at the conference examined how people learn from their neighbours and how influence spreads in a network, others looked at institutions that impose constraints on local interactions and who can link with whom. There was also a selection of papers offering insights on the role of social networks in influencing bargaining, besides the disruption of networks caused by the exit of players. My favourite paper was the one on how people bond together when facing a common enemy.
Ottoman models
But even better, if not the best thing, that the conference had to offer was Istanbul itself.
Spread across two continents, this is a city steeped in history and culture. The sweltering July heat during the day was, no doubt, exhausting, but as the folks here say, when the sun sets, even the most harassed traveller will forget all the day’s misadventures and embrace the city with open arms.
The cool sea breeze, night views of the Bosporus strait, the mezz e or the appetizer plate with numerous hot and cold offerings, fresh seafood under the Galata Bridge, endless varieties of local kebabs, raki (Turkish arrack) and mouth-watering desserts — I can fill an entire page with just that. Besides the Bosporus that separates the Asian and European parts of the city, there is also an estuary called the Golden Horn dividing old and new Istanbul. The Bosporus and the Golden Horn converge to the Sea of Marmara, on which a bevy of boats, ferries and ships — large, small and just as numerous as the fish in the water — ride.
Old Istanbul reminds you of Old Delhi with stunning examples of Ottoman architecture — the magnificent Topkapi Palace, the Sultan Ahmed mosque with its beautiful blue tiles, the Hagia Sophia (a church that became a mosque), the New Mosque (that was actually completed in 1665) and the Grand Bazaar, which is somewhat like our own Chandni Chowk.
While the typical Ottoman architecture is a blend of Byzantine and Islamic styles, the one I really liked was the Dolmabahce Palace, built on the Bosporus in the latter half of the Ottoman rule, incorporating both Baroque and Rococo elements.
As you take the guided tour through the palace, you are stunned not only by its gilded splendour, but also the constant footfalls and whispers that you almost hear in the corridors.
In the hot, heavy and still July air, you get a feel of the conspiracies and court intrigues that a dying empire may have witnessed.
Quite a lot of game theory there.
(The author is Professor of Economics, specialising in industrial organisation and applied game theory, at Louisiana State University.)