Re: Order of CASVB orbitals

Posted by Thorstein Thorsteinsson on February 14, 2002 at 09:58:11:

In Reply to: Order of CASVB orbitals posted by Takao Kobayashi on February 14, 2002 at 09:16:45:

: I would like to change the order (or numbering) of the orbitals in the CASVB
: wavefunction as I like, because I would like to know the weights of the preferable
: structures in the CASVB wavefunction. I have tried to change the orders by using
: Keywords "ORBPERM" and the order has somewhat changed. However I could not
: have obtained the order which I expect. Please teach me how to change the order
: of the orbitals in the CASVB wavefunction as I like.

ORBPERM should do what you want. The idea is that it permutes orbitals and makes the
corresponding changes to the structure coefficients. But note that the behaviour depends
on where ORBPERM is put in the input. If you, for example, put it in the beginning, the
(semi-random) starting-guess orbitals will be permuted. The converged orbitals may hence
not be what you expect. If you converge the orbitals and *then* use ORBPERM, it should
work the way you expect.

Another potential source of confusion is that several ways of defining permutations
exist. In casvb, ORBERM,3,-1,2,4 means [orbital 1 replaced by old orbital 3], [2 by -1],
[3 by 2], and [4 by 4].

: In addition, I have failed in taking the VB results from the VB information file
: (VBWFN) using Keyword "START". Please show me how to read the VB results
: from VBWFN.

Something like
START
VB=JOBOLD
should work. Can you give a little more details about how your input looks like?

Follow Ups:

• Re: Order of CASVB orbitals Takao Kobayashi 06:11:03 2/15/02 (1)
  • Re: Order of CASVB orbitals Thorstein Thorsteinsson 14:24:03 2/15/02 (0) 213.129.30.130

Post a Followup

Name:
E-Mail:

Subject:

Comments:
: : I would like to change the order (or numbering) of the orbitals in the CASVB : : wavefunction as I like, because I would like to know the weights of the preferable : : structures in the CASVB wavefunction. I have tried to change the orders by using : : Keywords "ORBPERM" and the order has somewhat changed. However I could not : : have obtained the order which I expect. Please teach me how to change the order : : of the orbitals in the CASVB wavefunction as I like. : ORBPERM should do what you want. The idea is that it permutes orbitals and makes the : corresponding changes to the structure coefficients. But note that the behaviour depends : on where ORBPERM is put in the input. If you, for example, put it in the beginning, the : (semi-random) starting-guess orbitals will be permuted. The converged orbitals may hence : not be what you expect. If you converge the orbitals and *then* use ORBPERM, it should : work the way you expect. : Another potential source of confusion is that several ways of defining permutations : exist. In casvb, ORBERM,3,-1,2,4 means [orbital 1 replaced by old orbital 3], [2 by -1], : [3 by 2], and [4 by 4]. : : In addition, I have failed in taking the VB results from the VB information file : : (VBWFN) using Keyword "START". Please show me how to read the VB results : : from VBWFN. : Something like : START : VB=JOBOLD : should work. Can you give a little more details about how your input looks like?