Stochastic simulation of sequential game-theory voting models

摘要

We discuss the use of stochastic simulation as a tool to learn about optimal behavior and Nash equilibria of a sequential voting model proposed by Osborne, related to Duverger's law. We introduce a graphical Java applet, which implements such simulations and investigates its properties. We show that in an appropriate setup, the applet is guaranteed to eventually find behavior, which is within epsilon of being optimal.