‘Learning to play’

Andreu Mas-Colell, Fabrizio Germano, Gábor Lugosi, Jens Josephson.

The main aim of this project is to analyse the different learning forms adopted by players. What decisions do players take in the medium and/or the long term? Do these decisions coincide with a particular Nash equilibrium or a particular correlated equilibrium? In what kinds of games and under what informational conditions can players learn to act as in a sequential equilibrium?

A lot of the research done to date concerns games in a normal form rather than degenerate games. We wish to explore the specific adaptations of learning processes made in the course of extensive games. Another study topic is how players learn to make decisions in an environment changing gradually in time.