Swap the end-points of two Markov chains

This is a conditional swap, according to the rules of parallel tempering. This function is only useful if the Markov chains have returned to the global scope and one process will make the decision and perform the swap, i.e.: the current state of each chain is locally available.

Usage

swap_points_locally(parMCMC)

Arguments

parMCMC

a list of Markov chain end points, each entry annotated with a temperature attribute attr(parMCMC[[i]],"beta")

Value

a list with some members swapped