|
Importance Sampling
in Monte Carlo Simulation of Rare Transition Events
|
The purpose of these three lectures is to bring out the
inner connections among the importance sampling method,
Monte Carlo simulation of random walkers, and rare
transition events between metastable states. These
connections lead to a new way of boosting the efficiency of
atomistic simulations whose time scale is severely limited
due to rare transition events. The discussion is focused on
computational methods with illustrations on low-dimensional
model problems. The materials are roughly organized as the
following:
(1) Monte Carlo computation of integrals and importance sampling as a
variance reduction technique.
(2) Using random walks to solve integral equations and partial
differential equations, and application of importance sampling.
(3) Using importance sampling to boost efficiency of Monte Carlo
simulations of rare transition events.
|
Meeting Time & Place
10:30-11:30am, August 1-3, 2005, B219, R163, LLNL
|
|