Event Type:

Department Colloquium

Date/Time:

Monday, November 22, 2010 - 08:00

Location:

Kidder 350

Local Speaker:

Abstract:

In this talk I will present an example of bounding total variation distance to stationary state and estimating mixing times via orthogonal polynomials diagonalization of discrete reversible Markov chains. Next, I will compare the orthogonal polynomials approach to the other known techniques, such as geometric convergence and coupling. I will proceed by suggesting a method for estimating mixing rates for certain examples of reversible Markov chains over general state spaces.