11 February 2011, Computational Social Choice Seminar, Vincent Conitzer
Game-theoretic analyses of strategic voting are generally troubled by equilibrium selection problems. In this talk, I will discuss our work on two complete-information models (where the preferences are common knowledge among the voters) that lead to a unique outcome. In one, the voters vote in sequence; in the other, there are multiple binary issues that are sequentially voted on by the voters. We consider a number of related questions, including how to compute the solution (involving a relation to compilation complexity), paradoxes, comparisons to truthful voting, and communication complexity.
This is joint work with Lirong Xia and Jérôme Lang.