An Isaac Newton Institute Workshop

Recent Advances in Monte Carlo Based Inference

Configurationally-Biased MC and Virtual-move parallel tempering

3rd November 2006

Authors: D. Frenkel (*) (FOM Institute AMOLF), I. Coluzza (Dept of Chemistry, U Cambridge, UK)


Parallel tempering is a Monte Carlo scheme that was introduced by Geyer in 1991 as a tool to boost the efficiency of Monte Carlo sampling Mendelian genetics. The technique has found widespread application in statistical mechanics. I will describe how, in some cases, the efficiency of the method may be increased by including information about MC trial moves that are excluded from the Markov chain. If time allows, I will discuss Biased Monte Carlo schemes that allow us to sample the conformations of composite objects, such as polymers.