English

Systematic enumeration of configuration classes for entropic sampling of Ising models

Statistical Mechanics 2011-12-07 v1 Computational Physics

Abstract

We describe a systematic method for complete enumeration of configuration classes (CCs) of the spin-1/2 Ising model in the energy-magnetization plane. This technique is applied to the antiferromagnetic Ising model in an external magnetic field on the square lattice, which is simulated using the tomographic entropic sampling algorithm. We estimate the number of configurations, Ω(E,m,L)\Omega(E,m,L), and related microcanonical averages, for all allowed energies EE and magnetizations mm for L=10L = 10 to 30, with ΔL=2\Delta L = 2. With prior knowledge of the CCs, we can be sure that all allowed classes are sampled during the simulation. Complete enumeration of CCs also enables us to use the final estimate of Ω(E,m,L)\Omega(E,m,L) to obtain good initial estimates, Ω0(E,m,L)\Omega_0(E,m,L'), for successive system sizes (L>LL' > L) through a two-dimensional interpolation. Using these results we calculate canonical averages of the thermodynamic quantities of interest as continuous functions of temperature TT and external field hh. In addition, we determine the critical line in the hh-TT plane using finite-size scaling analysis, and compare these results with several approximate theoretical expressions.

Keywords

Cite

@article{arxiv.1112.1387,
  title  = {Systematic enumeration of configuration classes for entropic sampling of Ising models},
  author = {Bruno Jeferson Lourenço and Ronald Dickman},
  journal= {arXiv preprint arXiv:1112.1387},
  year   = {2011}
}