Logit Dynamics for Local Interaction Games

Logit Dynamics for Local Interaction Games

Seminar author:Diodato Ferraioli

Event date and time:09/16/2015 02:30:pm

Event location:IFAE seminar room

Event contact:

The logit choice function is a family of randomized best response functions
parametrised by beta, the inverse noise level, which is used to model
players with limited rationality and knowledge [D. McFadden – Frontiers in
Econometrics, 1974]. We study the behavior of a game when players update
their strategies according to the logit choice function. We focus on two
extremal case: when at each step only one randomly chosen player is allowed
to update and when at each time step players concurrently update.
We study properties of these dynamics mainly in the context of local
interaction games, a class of games that has been used to model complex
social phenomena, including the spread of information and norms in social
networks, and physical systems, like the Ising model for spin systems. In a
local interaction game, the players are the vertices of a social graph and
the edges are two-player potential games. Each player picks one strategy to
be played for all the games she is involved with and the payoff of the
player is the (weighted) sum of the payoffs from each of the games.