Noisy heteroclinic networks and sequential decision making.

Math Physics Seminar
Wednesday, April 13, 2011 - 16:30
1 hour (actually 50 minutes)
Skiles 006
Georgia Tech
I will talk about sequential decision making models based ondiffusion along heteroclinic networks of dynamical systems, i.e.,multiple saddle-type equilibrium points connected by heteroclinicorbits. The goal is to give a precise description of the asymptoticbehavior in the limit of vanishing noise.In particular, I will interpret exit times for stochastic dynamics asdecision making times and give a result on their asymptotic behavior.I will report on extensive data on decision making in no a priori biassetting obtained in a psychology experiment that I ran with JoshuaCorrell (University of Chicago),and compare the data with my theoretical results. I will also showthat the same kind of limiting distribution for exit times appears innonequilibrium models of statistical mechanics.