Closed sethaxen closed 7 years ago
SerialMover enables a Gibbs Sampling scheme, where particles (ie, XYZ, RigidBodies, Scales) are all moved once. Nowhere is strictly recommending this mover with incremental scoring function, or at least I don't quite understand your question.
Agree with Riccardo, given a list of movers (eg. individual bead movers,rigid body movers..) it enables performing an MC step once per mover in the order specified in the list (Gibbs sampling essentially). The SerialMover only specifies how the Gibbs sampling is done and doesn't impose any requirements on the scoring function.
On Fri, Mar 17, 2017 at 3:50 AM, Riccardo Pellarin <notifications@github.com
wrote:
SerialMover enables a Gibbs Sampling scheme, where particles (ie, XYZ, RigidBodies, Scales) are all moved once. Nowhere is strictly recommending this mover with incremental scoring function, or at least I don't quite understand your question.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/salilab/pmi/issues/229#issuecomment-287323852, or mute the thread https://github.com/notifications/unsubscribe-auth/AH7fbxSy-ZhD-_kX--goFD7T_rps7Zpjks5rmmVZgaJpZM4MgTGv .
Should we close this?
Yes, thanks, you guys are absolutely right.
Perhaps I'm wrong, but it looks like
IMP.core.SerialMover
is used inIMP.pmi.samplers.MonteCarlo
to propose movers in order, one for each Monte Carlo step. My understanding is that this mover is only recommended when using an incremental scoring function.