Content | Navigation |

Physics

 


Seminars

Fall 2017 Seminars

DateTitleAuthor
August 28, 2017Methods for cyclic coating test protocol developmentAaron Feickert
September 4, 2017

 <Labor day>

September 11, 2017Molecular Dynamics Lattice Gas: a New Tool to Understand Lattice Gases and Lattice BoltzmannAlexander Wagner
September 18, 2017Differential Capacitance of Ionic Liquids with and without Added SolventSylvio May
September 25, 2017Probing student reasoning approaches through the lens of dual-process theoriesCody Gette
October 2, 2017Lattice Gas with a Molecular Dynamics collision operatorReza Parsa
October 9, 2017Rejection-free cluster Monte Carlo algorithms in the presence of symmetry-breaking fieldsYen Lee Loh
October 16, 2017
October 23, 2017
October 30, 2017
November 6, 2017<TBA>Mike Kessler
November  13, 2017
November 20, 2017
November 27, 2017<TBA>Andrew Croll
December 4, 2017
December 11, 2017<Final Exam Week>

 

Rejection-free cluster Monte Carlo algorithms in the presence of symmetry-breaking fields

Yen Lee Loh* and Nathan Carlson**

*Associate Professor
Department of Physics
University of North Dakota
**Graduate Student
Department of Physics
University of North Dakota

 

Monday October 9, 3:00-4:00pm, Refreshments at 2:30.

221 South Engineering

Cluster Monte Carlo algorithms are extremely efficient methods for generating thermal equilibrium configurations of spin models, lattice gases, and other systems.  However, most cluster algorithms rely on self-inverse global symmetries.  For models with symmetry-breaking external fields, the approach in the literature is to augment the cluster algorithm with a Metropolis criterion, which produces large rejection ratios and kills efficiency.  We have recently developed two cluster algorithms that have a 100% acceptance ratio even in the presence of symmetry-breaking fields.   The auxiliary-spin cluster algorithm (ASCA) replaces the symmetry-breaking field by a symmetry-preserving coupling to an auxiliary spin.   Our tests indicate that ASCA is superior to existing algorithms for Ising ferromagnets in uniform field.   The replica-exchange cluster algorithm (RECA) evolves two copies of the system using “swaps” (cluster-exchange updates), interleaved with “rotations” (global symmetry operations that need not be self-inverse), and with other types of moves (such as Metropolis updates) to ensure ergodicity.  It is applicable to XY models in the presence of vector potentials.   A Monte Carlo algorithm combining ASCA, RECA, and Metropolis updates may be a promising way to simulate various systems including superconducting Josephson junction arrays in magnetic fields, frustrated magnets, magnetic skyrmions, and molecular fluids.  We are currently benchmarking the algorithms to compare their autocorrelation times with other Monte Carlo methods.

o either subscribe to or unsubscribe from this listserve please send an email to Listserv@listserv.nodak.edu with an empty subject and containing a body of
SUB NDSU-PHYSICS-SEMINAR or SIGNOFF NDSU-PHYSICS-SEMINAR respectively.


Student Focused. Land Grant. Research University.

Follow NDSU
  • Facebook
  • Twitter
  • RSS
  • Google Maps

Site Manager: Physics Webmaster
Published by Department of Physics

Last Updated: Friday, October 06, 2017 5:22:35 PM
Privacy Statement