[Wien] Question: geometry optimization: constrains

[L. D. Marks]

There is a general method which should work for the PORT option. What 
you will need to define are the matrices T_1, which transforms from the 
full co-ordinates to some reduced set, a matrix T_2 which will transform 
the Cholesky of the Hessian from the full co-ordinates to some reduced set 
and their inverses TI_1, TI_2. You would then bracket the call to DRMNG on 
line 145 by the relevant transformations, i.e. (using * for a matrix 
multiplication)
 	GNEW = T_1 * G
 	XNew = T_1 * X
 	HNEW = T_2 * HESS
         call DRMNG(D, FX, GNEW, IV, LIV, LV, N, V, XNEW, HNEW)
 	X = TI_1 * XNEW
 	HESS = TI_2 * HNEW

There might also be a change in the number of parameters N. It should not 
be too hard to write something relatively general for this -- I can 
provide more information off-line if needed. There will be at least one 
improvement to the PORT option in the near future (speedup of 1/3 to 2/3), 
so it would be a good time to merge something like this in if someone is 
interested in contributing some time.

On Wed, 3 May 2006, Oleg Rubel wrote:

> Dear WIEN2k Users,
>
> could you please comment regarding to setting of constrains when
> performing a geometry optimization with WIEN2k.
>
> The constrains for geometry optimizations (internal degrees of freedom)
> are specified in file case.inM.
>
> ---- top of the file: case.inM -----
> PORT 2.0           # PORT/NEWT/NEW1;  tolf (geom.opt...
> 1.0 1.0 1.0 1.0    # Atom1
> ...
> ---- bottom of the file: case.inM -----
>
> As far as I understand, this constrains correspond to the global
> coordinates and, if setting them to 0.0, allow one to restrict the motion
> of atoms along X, Y, and Z global (!) axes.
>
> QUESTION: What to do if I would like to constrain positions of an
> atom to movement in a plain [say (111) plane for the primitive
> cubic symmetry], which is NOT parallel to XY, XZ, or YZ planes of
> the global coordinate system? Such an option is required for
> simulation of annealing processes, when an atom moves from one
> local minima to another along a certain path, which needs to be
> determined.
>
>
> Thank you in advance.
>
> Oleg Rubel
>
> ===========================
> Dr. Oleg Rubel
> Semiconductor Physics Group
> and Material Science Center
> Department of Physics
> Philipps University Marburg
> Renthof 5, D-35032 Marburg, Germany
> Tel.: +49-(0)6421-2825726
>      +49-(0)6421-2824241
> Fax.: +49-(0)6421-2828935
>      +49-(0)6421-2827076
> E-mail: Oleg.Rubel at physik.uni-marburg.de
> Homepage: http://www.staff.uni-marburg.de/~rubel/
>
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>

Note: if you have an old email address for me, please note that "nwu" has
been changed to "northwestern".
-----------------------------------------------
Laurence Marks
Department of Materials Science and Engineering
MSE Rm 2036 Cook Hall
2225 N Campus Drive
Northwestern University
Evanston, IL 60208, USA
Tel: (847) 491-3996 Fax: (847) 491-7820
email: L-marks at northwestern dot edu
http://www.numis.northwestern.edu
-----------------------------------------------