Algorithmic Game Theory

January 10 - 14, 2011

Overview

The wealth of strategic interactions among Internet agents with very diverse interests, in varying degrees of competition and cooperation, naturally calls for a fusion of tools from computer science, game theory and economics. A new research area called Algorithmic Game Theory (AGT) has emerged as a result of such a fusion. However, AGT is not just about applying analytical tools from computer science to game theory/economics or vice versa but primarily about providing new conceptual perspectives at a very fundamental level. For example, while efficient computability of best responses is a natural consideration for an algorithm designer, this in turn lends questionability to the classical economics assumption that agents be fully rational, opening way for study of equilibria under bounded rationality. On the other hand, the desirable game-theoretic solution concepts (e.g. correlated, Nash, truthfulness) for different scenarios must be fundamentally guided by underlying economic principles, and already a richness can be seen in the feedback between the two fields.

While Nash equilibria are difficult to compute even for two player games, there exists an efficient algorithm to compute correlated equilibria for any succinctly representable game. An example of a more lucrative interplay that involves the design of mechanisms guaranteed to satisfy some desirable stability requirements (e.g. truthfulness) all the while also minimizing overpayments has opened way to a boom of sponsored search advertising that dominates much of the internet industry today. Indeed, the scope and diversity of the Internet economy and the social transactions that can be potentially studied and analyzed via algorithmic game theoretic techniques has been exploding exponentially, and there is a need for continued dialogs among the various communities to get a better understanding of the underlying concepts and issues.

This workshop will gather scientists and researchers from various communities such as mathematics, computer science, economics, game theory, information theory, from academia as well as industry to provide a joint platform to discuss fundamental issues in the emerging interdisciplinary field of Algorithmic Game Theory.

Organizing Committee

Gunes Ercal (University of Kansas)
Allon Percus (Claremont Graduate University)
Vwani Roychowdhury (University of California, Los Angeles (UCLA))
Sudhir Singh (University of California, Los Angeles (UCLA))