On Thu, Feb 21, 2008 at 7:11 AM, Matthew Toseland <toad at amphibian.dyndns.org> wrote: > So a swap will only happen if it minimises stress for *both* nodes, rather > than before it would happen if it minimised total stress for both nodes > combined. Clearly this is a significant change to the transition > probabilities... It might very well lead to false minima? One thing to try is > to use the same random for each node's decision (since this is likely how it > will be implemented in practice).
I really don't think we should be making this kind of change without Oskar's agreement, since it is very possible that this could render the underlying principals behind the Metropolis-Hastings algorithm invalid. One of the nice things about the current swap algorithm is that it has a solid mathematical foundation, we can't make changes which throw that away just on a whim. Ian. -- Email: ian at uprizer.com Cell: +1 512 422 3588 Skype: sanity
