Geometric allocation approaches in Markov chain Monte Carlo

ELC International Meeting on Inference, Computation, and Spin Glasses

Todo, S.^1 ; Suwa, H.^2

Institute for Solid State Physics, University of Tokyo, 7-1-26-R501 Port Island South, Kobe 650-0047, Japan^1

Department of Physics, Boston University, 590 Commonwealth Avenue, Boston, MA 02215, United States^2

关键词: Allocation approach;    Computational time;    Long range interactions;    Markov chain Monte Carlo method;    Markov Chain Monte-Carlo;    Multi dimensional;    Transition probabilities;    Unconventional approaches;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/473/1/012013/pdf
DOI  :  10.1088/1742-6596/473/1/012013

来源: IOP

【 摘 要 】

The Markov chain Monte Carlo method is a versatile tool in statistical physics to evaluate multi-dimensional integrals numerically. For the method to work effectively, we must consider the following key issues: the choice of ensemble, the selection of candidate states, the optimization of transition kernel, algorithm for choosing a configuration according to the transition probabilities. We show that the unconventional approaches based on the geometric allocation of probabilities or weights can improve the dynamics and scaling of the Monte Carlo simulation in several aspects. Particularly, the approach using the irreversible kernel can reduce or sometimes completely eliminate the rejection of trial move in the Markov chain. We also discuss how the space-time interchange technique together with Walker's method of aliases can reduce the computational time especially for the case where the number of candidates is large, such as models with long-range interactions.