Towards optimal scaling of metropolis-coupled Markov chain Monte Carlo
Towards optimal scaling of metropolis-coupled Markov chain Monte Carlo
复制标题
DOI：
10.1007/s11222-010-9192-1
复制
发表时间：
2011-10-01
影响因子：
2.2
通讯作者：
Rosenthal, Jeffrey S.
中科院分区：
文献类型：
--
作者：
Atchade, Yves F.;Roberts, Gareth O.;Rosenthal, Jeffrey S.
我们认为最佳的温度间隔的大都市耦合马尔可夫链蒙特卡罗（MCMCMCMC）和模拟回火算法。我们证明，在某些条件下，这是最佳的（在最大化的预期平方跳跃距离），以空间的温度，使被接受的温度互换的比例约为0.234。这概括了物理学家的相关工作，并与以前的工作是一致的随机行走大都会算法的最佳缩放。
We consider optimal temperature spacings for Metropolis-coupled Markov chain Monte Carlo (MCMCMC) and Simulated Tempering algorithms. We prove that, under certain conditions, it is optimal (in terms of maximising the expected squared jumping distance) to space the temperatures so that the proportion of temperature swaps which are accepted is approximately 0.234. This generalises related work by physicists, and is consistent with previous work about optimal scaling of random-walk Metropolis algorithms.