Table of Contents

2023/11/14 18:37

A new coupling technique

Let , be two probability measures on the same measurable space .

We draw jointly the couple of random variables according to the following procedure:

Proposition. is a coupling of .

  • What is nice is that we are able to couple these random variables whereas their densities are known only up to a multiplicative constant. I wonder if it is interesting to couple in that way: and . Up to some tricks, can we deduce a way of coupling two MH starting from different initial distributions, maybe with delayed coupling? Can we compare it to the coupling of Pierre Jacob et al.?
  • Moreover, if we look at the problem today, if then, is a coupling of , no?

Proof

Obviously, where

We now show that is a coupling of . To do so, it is sufficient to check that for any bounded or non-negative function , .

Indeed, write, using the detailed balance condition in the second line : which completes the proof.

Some comments

The coupling probability is given by:

Question: we know that . But I can't see how to prove