Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

DY: Fachverband Dynamik und Statistische Physik

DY 50: Extreme Events (joint session DY/ SOE)

DY 50.2: Vortrag

Donnerstag, 19. März 2015, 15:30–15:45, BH-N 243

Computing the probability of rare trajectories — •Jorge C. Leitão¹, João M. Viana Parente Lopes², and Eduardo G. Altmann¹ — ¹Max Planck Institute for the Physics of Complex Systems, 01187 Dresden, Germany — ²Department of Physics and Center of Physics, University of Minho, P-4710-057, Braga, Portugal

Estimating the probability of extreme events often requires finding rare trajectories in (high-dimensional non-linear) dynamical systems. In this talk we show how such trajectories can be efficiently sampled using importance sampling Monte Carlo methods. We argue that the applicability and efficiency of these methods depend on the sensitivity of the observed quantity (in which the extreme event is measured) to perturbations of the initial conditions (in the phase space). We show analytical results and numerical simulations for different observables in (hyperbolic and non-hyperbolic) chaotic systems.

- J. C. Leitão, J. M. Viana Parente Lopes, E. G. Altmann "Efficiency of Monte Carlo Sampling in Chaotic Systems", Phys. Rev. E 90, 052916 (2014)

- J. C. Leitão, J. M. Viana Parente Lopes, E. G. Altmann "Monte Carlo Sampling in Fractal Landscapes", Phys. Rev. Lett. 110, 220601 (2013)