Posted by Per Åke Malmqvist on February 07, 2008 at 16:27:42:
In Reply to: Freezing CI coefficients posted by David Robinson on February 06, 2008 at 16:24:05:
: Dear Molcas users,
: Is it possible to freeze the CI coefficients for a few iterations of the orbital optimization?
Hi David,
No, this is not technically possible right now, but I do not quite understand the
purpose. Keeping the CI coefficients unchanged after an orbital optimization step
and continuing would simply mean to reassume orbital optimization in the next
iteration, and (unless the previous orbital optimization failed) there would
simply be nothing to do. The optimizer would find the orbital gradient (the BLB
elements) to be (almost) zero, and that would be it.
But if I misunderstand your purpose, please write back!
Best,
Per Åke