 
Summary: COLLOQUIUM
University of Regina
Department of Mathematics and Statistics
Speaker: Dr. Christopher Bose
University of Victoria, Canada.
Title: Maximum Entropy Methods in Ergodic Theory
Date: Tuesday, June 19, 2007
Time: 11:00 a.m.
Place: Math & Stats Lounge (CW 307.20)
Abstract
Ergodic Theory aims to study the statistical properties of a dynamical
system (X, T) with respect to an INVARIANT MEASURE. Some systems
come with a natural invariant measure (eg: Hamiltonian systems, Markov
chains etc) while others do not (eg: the family of Logistic maps on [0, 1]).
In this talk I will describe two methods for constructing invariant measures
in this latter situation.
The first is based on linear algebra. It is highlyeffective on a restricted
class of transformations acting on ddimensional Euclidean space. It is also
computationally efficient.
The second method is based on convex optimization. It is nearly univer
