Thursday 08 Feb 2018: Dynamics of Learning in Game Theory
Sofia Castro - University of Porto
LSI seminar room A 13:30-14:30
In Game Theory, decisions are made in order to optimise payoff. When a sequence of decisions occurs over time, these can change according to what is considered more favourable at a given decision time. The change comes from an update of beliefs and constitutes a learning mechanism which can be modelled by a dynamical system. Depending on the update procedure, different learning mechanisms are constructed.
I shall focus on two classic learning mechanisms in continuous time: replicator dynamics (RD) and best-response dynamics (BRD). It is known that the Nash equilibria, describing the outcome of the game, are the same for these two learning mechanisms. However, non-equilibrium solutions (along which learning is involved) can behave quite differently under RD and BRD. I shall address the relation between the solutions under RD and BRD from two points of view: those starting from the same initial state and those reaching the same limit set.