Slowdown estimates for random walk in random environment
Vortrag von Prof. Dr. Noam Berger Datum: 20.04.09 Zeit: 14.00 - 00.00 Raum: Y27H25
Abstract: Random walk in random environment, introduced by Solomon in 1975, is
one of the most
studied models for transport in irregular media. It constitutes a
Markov chain on the lattice
Z^d, where the transition probabilities are determined in some random
manner, whose distribution
is typically translation invariant and ergodic.
For a class of models of random walk in random environment,
we discuss the following problem: What is the probability that the
particle moves
at a speed significantly slower than its predicted asymptotic speed
(to be explicit,
we say less than half the predicted asymptotic speed)?
As it turns out, it is easy to give a lower bound for this
probability, based on the so called
"naive trap analysis". In dimension greater than or equal to 4, we
show an upper bound
for this probability which is very close to the lower bound obtained
by the naive trap analysis.
This improves previously known bounds.
As a tool for obtaining the main result, we also improve certain
central limit theorem estimates
for this process.
No prior familiarity with random walk in random environment will be
assumed. All necessary
concepts will be explained during the talk.