|
The
Summer Session is a 3-week residential program in Park
City, Utah and is part of the larger PCMI program.
Application Deadline
February 15, 2007
|
|
|
The PCMI 2007 Program
Undergraduate Faculty Program (UFP)
Co-sponsored by Chautauqua Programs
For the faculty members whose main focus is teaching undergraduate students,
PCMI offers the opportunity to renew excitement about mathematics, talk
with peers about new teaching approaches, address some challenging research
questions, and interact with the broader mathematical community. Each
year the theme of the UFP bridges the research and education themes of
the Summer Session.
MARKOV CHAINS AND MIXING TIMES
David Levin, University of Oregon
Over the past twenty years, research in discrete probability – and finite Markov chains in particular – has enjoyed a great renaissance. The
aim of the Undergraduate Faculty Program (UFP) is to communicate
how these recent research developments can invigorate the undergraduate
probability curriculum.
As part of the UFP, we will present the rudiments of a course, based
on the forthcoming book Markov Chains and Mixing by D. Levin, Y.
Peres, and E. Wilmer, on bounding the convergence time for finite
Markov chains. (See below for a description of this topic.) Many of the
examples presented will come from statistical echanics, the topic area
of the Institute this year. This course will serve as a model for a second
college-level course in probability, and the only background required of
participants is an undergraduate course in probability. Lectures will
be supplemented by discussions and examples of interactive exercises
which might be incorporated into a college-level course.
About Mixing Times
How many times must a deck of cards be shuffled before it is wellmixed?
If a walker explores a network by moving at each step to
a randomly chosen neighboring location, how many steps must she
take before the distribution of her location is spread over the entire
network with little dependence on her starting position? These are
both examples of Markov chains, randomly evolving systems which
under mild assumptions settle down into steady-state distributions.
The time required for a Markov chain to converge to its equilibrium
distribution is is called the mixing time, and estimating mixing times
is an active area of research in probability theory. During the UFP,
we will show how exciting developments in the study of mixing times – with connections to physics and computer science – can be given an
elementary exposition accessible to motivated undergraduates.
The Coordinator of PCMI's Undergraduate Faculty Program is William Barker, Bowdoin College. |