Astrophysics (Index)About

umbrella sampling

(method usable in MCMC that avoids some problems)

Umbrella sampling is a method used within Markov chain Monte Carlo (MCMC) to chose the values to evaluate, that avoids some issues that can occur with the more-usual, earlier-developed Metropolis algorithm and its variants. The Metropolis algorithm does not sufficiently cover the parameter space of some functions, and Umbrella sampling is specifically designed to overcome this.


(technique,models,mathematics,computation)
Further reading:
https://en.wikipedia.org/wiki/Umbrella_sampling
https://ui.adsabs.harvard.edu/abs/2018MNRAS.480.4069M/abstract
https://ui.adsabs.harvard.edu/abs/1977JCoPh..23..187T/abstract

Index