Rates of convergence of Markov chains to stationarity : strong stationary times, coupling, Gelfand pairs and comparison theory

Placeholder Show Content

Abstract/Contents

Abstract
The main purpose of this thesis is to find the mixing time for various random walks that the math community has been interested in. The methods used in this thesis are strong stationary times, introduced by Aldous and Diaconis, a coupling argument that was inspired by work of Aldous, bounding the eigenvalues of a walk via comparison theory and path arguments introduced by Diaconis and Saloff-Coste and via Gelfand pair theory developped by Diaconis and Shahshahani, but also finding the eigenvalues of random walks using Fourier Transform arguments as introduced by Diaconis and Shahshahani. There are many norms that can use to study the mixing time of a random walk on a finite group. Coupling is used to bound the $l^1$ norm, strong stationary times are used to bound the separation distance, while the eigenvalues of a reversible Markov chain are used to bound the $l^2$ norm. This thesis focuses on specific examples of Markov chains on finite spaces, some of which have been studied before but through a different norm.

Description

Type of resource text
Form electronic; electronic resource; remote
Extent 1 online resource.
Publication date 2016
Issuance monographic
Language English

Creators/Contributors

Associated with Nestoridi, Evrydiki-Xenia
Associated with Stanford University, Department of Mathematics.
Primary advisor Diaconis, Persi
Thesis advisor Diaconis, Persi
Thesis advisor Soundararajan, Kannan, 1973-
Thesis advisor Vondrak, Ján, (Mathematician)
Advisor Soundararajan, Kannan, 1973-
Advisor Vondrak, Ján, (Mathematician)

Subjects

Genre Theses

Bibliographic information

Statement of responsibility Evrydiki-Xenia Nestoridi.
Note Submitted to the Department of Mathematics.
Thesis Thesis (Ph.D.)--Stanford University, 2016.
Location electronic resource

Access conditions

Copyright
© 2016 by Evrydiki-Xenia Nestoridi

Also listed in

Loading usage metrics...