Page 1 of 1

Indexes in correlation.data file

Posted: Sat Apr 29, 2017 5:24 pm
by Vladimir_Konjkov
Hi all.
My question is about correlation.data file.

I find detailed explanation about JASTRO factor used for finit system and placed in correlation.data file. https://arxiv.org/pdf/0801.0378.pdf
In formulas (19), (20), (21) alpha, beta, gamma have 1, 2 and 4 indexes respectively, but in correlation.data I have found next lines

7.5982746075803145E-003 1 ! alpha_0,1
1.3426533565298453E-002 1 ! alpha_0,2
-7.2944837829337049E-003 1 ! beta_0,1,1

-2.4907681898067393E-003 1 ! gamma_1,1,0,1,1
3.5449418331906443E-003 1 ! gamma_2,1,0,1,1
2.1243610924541823E-003 1 ! gamma_2,2,0,1,1
-3.2040038712909851E-004 1 ! gamma_0,0,2,1,1
-1.2600199356446610E-003 1 ! gamma_1,0,2,1,1
-1.2640600490430177E-003 1 ! gamma_1,1,2,1,1
5.9060512214081267E-004 1 ! gamma_2,1,2,1,1
What is the correspondence between the indices in formulas and file and what are the extra index for?

Thank you in advance for answering my simple question.

Vladimir.

Re: Indexes in correlation.data file

Posted: Sat Apr 29, 2017 5:57 pm
by Ryan_Hunt
Dear Vladimir,

The final index for alpha, and the second-to-last indices for beta and gamma are the "spin-pair-type" indices. If one allows for a spin dependence of the corresponding terms in the Jastrow exponent, then these indices may differ from each other (e.g. if you make the simple choice of setting spin dependence uu=dd=/=ud, you will find alpha terms with final indices of both "1" and "2", and of course there will be twice as many optimisable parameters in this case).

The final indices for beta and gamma are the "ion set". For example, if the gamma or beta coefficients appear in "SET 1" blocks, the final index will always be "1". If they appeared in a "SET 2" block, they will always read "2", etc. This is common if one studies a system made up of more than one kind of atom (where one would presumably want to specify different Jastrow terms for each!)

Regards,
Ryan.

Re: Indexes in correlation.data file

Posted: Sun Apr 30, 2017 2:26 am
by Vladimir_Konjkov
Thank you Ryan, your explanation was very helpful.