Coupling Control Variates for Markov Chain Monte Carlo
Numerical Analysis
2015-05-13 v1 Computational Physics
Computation
Abstract
We show that Markov couplings can be used to improve the accuracy of Markov chain Monte Carlo calculations in some situations where the steady-state probability distribution is not explicitly known. The technique generalizes the notion of control variates from classical Monte Carlo integration. We illustrate it using two models of nonequilibrium transport.
Cite
@article{arxiv.0809.2300,
title = {Coupling Control Variates for Markov Chain Monte Carlo},
author = {Jonathan B. Goodman and Kevin K. Lin},
journal= {arXiv preprint arXiv:0809.2300},
year = {2015}
}