Seminario Dottorato: An introduction to stochastic ergodic control

Wednesday 15 January 2014, h. 15:15, room 1BC45
Marco Cirant (Padova - Dip. Mat.)
"An introduction to stochastic ergodic control"

Abstract
In this talk we give an introduction to stochastic ergodic control problems, where an agent aims at minimizing a long-time average cost by controlling his own state.
We will show, through a toy example, the main features of the problem and how it is possible to produce an optimal control by solving a suitable elliptic nonlinear partial differential equation.
In the final part of the talk we will explore briefly how the minimization problem for a single agent can be considered more in general for a continuum of identical agents. This research field is called Mean Field Games and has attracted the experts' attention in the last ten years.