Simulation of large deviation functions using population dynamics

Abstract

In these notes we present a pedagogical account of the population dynamics methods recently introduced to simulate large deviation functions of dynamical observables in and out of equilibrium. After a brief introduction on large deviation functions and their simulations, we review the method of Giardinà \emph{et al.} for discrete time processes and that of Lecomte \emph{et al.} for the continuous time counterpart. Last we explain how these methods can be modified to handle static observables and extract information about intermediate times.

Proceedings of the 10th Granada Seminar on Computational and Statistical Physics