University of Illinois Urbana-Champaign

Eliminating critical slowing down in Monte Carlo calculations

Luehrmann, Mia Kerstin

This item is only available for download by members of the University of Illinois community. Students, faculty, and staff at the U of I may log in with your NetID and password to view the item. If you are trying to access an Illinois-restricted dissertation or thesis, you can request a copy through your library's Inter-Library Loan office or purchase a copy directly from ProQuest.

Permalink

https://hdl.handle.net/2142/19039

Description

Title
Eliminating critical slowing down in Monte Carlo calculations
Author(s)
Luehrmann, Mia Kerstin
Issue Date
1991
Doctoral Committee Chair(s)
Stack, John D.
Department of Study
Physics
Discipline
Physics
Degree Granting Institution
University of Illinois at Urbana-Champaign
Degree Name
Ph.D.
Degree Level
Dissertation
Keyword(s)
  • Physics
  • Elementary Particles
  • High Energy
Language
eng
Abstract
  • We examine methods to improve the major numerical difficulties in lattice field theory. Traditional Metropolis and heat bath Monte Carlo methods in lattice calculations break down whenever one tries to calculate thermodynamic quantities near critical points; this phenomenon is called Critical Slowing Down, (CSD). Recently, alternate methods have been proposed to shorten the relaxation time and thereby, reduce CSD. These methods all modify site-by-site Metropolis and heat bath Monte Carlo to operate on larger spacial scales. One of these newer techniques is to apply multigrid methods to site-by-site Monte Carlo algorithms; another is to stochastically determine clusters of sites on the lattice by simplifying the Hamiltonian until it is determinate.
  • We have applied both techniques to the Ising model and compared the relaxation time constants to those determined by site-by-site Monte Carlo methods and found that they are lower. However, even after we succeeded in vectorizing the algorithms, the computation time needed to calculate each sweep of the lattice is larger than that needed by the site-by-site Monte Carlo methods. The important quantity is the computation time needed to move from one independent configuration to another, which is the time needed to calculate each sweep of the lattice multiplied by the relaxation time constant. The net result of the Multigrid Monte Carlo method is that it is less efficient than regular Monte Carlo whenever the parameters of the algorithm are fixed such that the algorithm satisfies detailed balance. The net effect of the Stochastic Blocking method is an improvement compared to regular Monte Carlo when the coupling constant is close to the critical point. We believe the methods used here can be adapted to lattice gauge theory calculations.
Type of Resource
text
Permalink
http://hdl.handle.net/2142/19039
Copyright and License Information
Copyright 1991 Luehrmann, Mia Kerstin

Owning Collections

Graduate Dissertations and Theses at Illinois PRIMARY
Graduate Theses and Dissertations at Illinois
Dissertations and Theses - Physics
Dissertations in Physics