LONG-TERM BEHAVIOUR OF LARGE MECHANICAL SYSTEMS WITH RANDOM INITIAL DATA

R. KUPFERMAN,A. M. STUART,J. R. TERRY,P. F. TUPPER
DOI: https://doi.org/10.1142/s0219493702000571
2002-12-01
Stochastics and Dynamics
Abstract:We study the long-time behaviour of large systems of ordinary differential equations with random data. Our main focus is a Hamiltonian system which describes a distinguished particle attached to a large collection of heat bath particles by springs. In the limit where the size of the heat bath tends to infinity, the trajectory of the distinguished particle can be weakly approximated, on finite time intervals, by a Langevin stochastic differential equation. We examine the long-term behaviour of these trajectories, both analytically and numerically. We find ergodic behaviour manifest in both the long-time empirical measures and in the resulting auto-correlation functions.
statistics & probability
What problem does this paper attempt to address?