Re: Monte Carlo, Genetic Algorithms (...) for Powder Structure

Alan Hewat, ILL Grenoble ( (no email) )
Wed, 28 Jan 1998 15:59:19 +0200

It is unfortunate that the question of using "genetic" or other algorithms
to solve crystal structures (prior to Rietveld refinement) has been mixed
up with the "problem" of extracting peak intensities. It is not altogether
surprising that it is faster to compare a model to integrated intensities
than it is to compare it to an "order of magnitude" more profile intensities
(using "genetic" algorithms or whatever you favour right now). But it is
perhaps going a bit far to say:

wifd@isise.rl.ac.uk (Bill David) wrote 28 Jan 1998:
"...reasons that should be obvious - fitting to the extracted intensities (with
the associated covariance matrix) is mathematically equivalent to fitting to
the pattern."

This implies that you don't even need to do the Rietveld refinement (which I
suppose Bill did not mean to imply :-) Extracting intensities is just convenient
and fast, not equivalent nor essential in this case.

As Armel says: "determine the structure whatever the method ... then use
the Rietveld method for the final refinement".

Alan Hewat, ILL Grenoble, FRANCE <hewat@ill.fr> tel (33) 4.76.20.72.13
ftp://ftp.ill.fr/pub/dif fax (33) 4.76.48.39.06 http://www.ill.fr/dif/