Optimizing in a Strategic World: A Survey of Recent Research in Algorithmic Game
The goal of discrete optimization is to design systems with optimal or near-optimal performance. In the age of the Internet, however, we must take into account the fact that many of the users of our systems are driven by an economic goal, and interact with varying degrees of collaboration and competition. Moreover, the strategic nature of interactions in online dynamic marketplaces means that the roll-out of a new algorithm designed with the expectation of improved performance can end up degrading performance due to unanticipated responses by strategic users.
The field of algorithmic game theory addresses this issue, as well as a wide variety of other problems at the intersection of game theory, economics and computer science. In this talk, we survey recent research and open problems in this field.
Anna Karlin, University of Washington