CONVERGENCE OF THE WANG-LANDAU ALGORITHM
CONVERGENCE OF THE WANG-LANDAU ALGORITHM
复制标题
DOI：
10.1090/s0025-5718-2015-02952-4
复制
发表时间：
2015-09-01
影响因子：
2
通讯作者：
Stoltz, Gabriel
中科院分区：
文献类型：
--
作者：
Fort, Gersende;Jourdain, Benjamin;Stoltz, Gabriel
分析了Wang-Landau算法的收敛性。这种抽样方法属于一般的自适应重要性抽样策略，它利用沿选定的反应坐标的自由能作为偏差。这些算法对于提高马尔可夫链蒙特卡罗算法在亚稳态下的采样性能非常有帮助。证明了Wang-Landau算法的收敛性及其中心极限定理。
We analyze the convergence properties of the Wang-Landau algorithm. This sampling method belongs to the general class of adaptive importance sampling strategies which use the free energy along a chosen reaction coordinate as a bias. Such algorithms are very helpful to enhance the sampling properties of Markov Chain Monte Carlo algorithms, when the dynamics is metastable. We prove the convergence of the Wang-Landau algorithm and an associated central limit theorem.