分子动力学中重要性抽样与增强抽样方法的一些联系。
Some connections between importance sampling and enhanced sampling methods in molecular dynamics.
机构信息
Zuse Institut Berlin, Takustrasse 7, 14195 Berlin, Germany.
出版信息
J Chem Phys. 2017 Nov 21;147(19):194107. doi: 10.1063/1.4989495.
In molecular dynamics, enhanced sampling methods enable the collection of better statistics of rare events from a reference or target distribution. We show that a large class of these methods is based on the idea of importance sampling from mathematical statistics. We illustrate this connection by comparing the Hartmann-Schütte method for rare event simulation (J. Stat. Mech. Theor. Exp. 2012, P11004) and the Valsson-Parrinello method of variationally enhanced sampling [Phys. Rev. Lett. 113, 090601 (2014)]. We use this connection in order to discuss how recent results from the Monte Carlo methods literature can guide the development of enhanced sampling methods.
在分子动力学中，增强采样方法能够从参考或目标分布中更好地收集罕见事件的统计信息。我们表明，这些方法中的一大类是基于数理统计中的重要性抽样思想。我们通过比较 Hartmann-Schütte 方法（用于罕见事件模拟，J. Stat. Mech. Theor. Exp. 2012，P11004）和 Valsson-Parrinello 方法（变分增强采样，Phys. Rev. Lett. 113，090601（2014））来说明这种联系。我们利用这种联系来讨论最近来自蒙特卡罗方法文献的结果如何指导增强采样方法的发展。
Zotero 插件