Condensed Matter > Statistical Mechanics
[Submitted on 1 Jul 2013 (this version), latest version 17 Jun 2015 (v8)]
Title:Acceleration without Detailed Balance Condition from Perspective of Nonequilibrium Behavior
View PDFAbstract:Monte-Carlo method is a major technique to investigate equilibrium properties governed by the well-known Gibbs-Boltzmann distribution through an artificial relaxation process. The standard way is to realize the Gibbs-Boltzmann distribution under the detailed balance condition. Recent progress on a technique proposed by Suwa and Todo makes it possible to accelerate the relaxation process, while avoiding the detailed balance condition. Since the dynamical stochastic system implemented to perform the Monte-Carlo simulation does not possess an equilibrium state but a steady state, we must design the steady state to be equivalent to the desired distribution such as the Gibbs-Boltzmann one. In other words, we then exploit nonequilibrium behavior with some current to accelerate the relaxation to the desired distribution. Several evidence in the previous studies exists but we do not have an answer to a simple question how the absence of the detailed balance condition accelerates the convergence to the desired distribution. In the present study, in order to address this problem, we find several relations in the nonequilibrium process without equilibrium but with steady state. By use of the obtained relations, we discuss the mechanism of the fast convergence in the case without the detailed balance condition.
Submission history
From: Masayuki Ohzeki [view email][v1] Mon, 1 Jul 2013 16:42:49 UTC (10 KB)
[v2] Thu, 19 Dec 2013 02:19:08 UTC (12 KB)
[v3] Thu, 3 Jul 2014 00:55:30 UTC (351 KB)
[v4] Thu, 25 Sep 2014 04:14:35 UTC (854 KB)
[v5] Wed, 5 Nov 2014 06:31:56 UTC (1,063 KB)
[v6] Thu, 13 Nov 2014 00:44:25 UTC (1,012 KB)
[v7] Mon, 16 Mar 2015 02:30:23 UTC (302 KB)
[v8] Wed, 17 Jun 2015 00:06:36 UTC (303 KB)
Current browse context:
cond-mat.stat-mech
Change to browse by:
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
IArxiv Recommender
(What is IArxiv?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.