Re: CCL:optimizer/Transition States

[Didier MATHIEU <mathieu (- at -) ripault.cea.fr>]

Dear Peter
 > The idea is to interface the optimizer with a different program that
 > I am using here in that I would require the source code
 > (Fortran would be best).
 Wouldn't the popular book "Numerical Recipes in Fortran" (Cambridge
 University Press, by Press, Flannery et al.) match your need ? I used
 routines of this book to perform local geometry optimizations and
 simulated annealing in cartesian coordinates, including constraints.
 To use a global search algorithm in internal coordinates, the main
 difference, to me, is to care with the differential volume element in
 internal coordinates if one wants to ensure uniform sampling in
 cartesian space.
 > Different optimization routines/algorithms were obviously a plus.
 I am specially concerned with OPTIMIZATION ALGORITHMS FOR TRANSITION
 STATES. I know those implemented in Gaussian and Ampac, but would be
 interested in getting a more exhaustive account (any review articles?)
 about all such algorithms.
 Regards
 --
 Didier MATHIEU
 CEA - Le Ripault, BP 16
 37260 Monts (France)
 Tel. 33(0)2.47.34.41.85