Approximating one-matrix functionals without generalized Pauli constraints

Oleg V. Gritsenko, Katarzyna Pernal

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

Common features are identified in the properties of the pure (PF) and ensemble (EF) energy functionals defined within the density matrix functional theory (DMFT) as well as their free-form (FFF) approximations. These features rationalize abandoning, in the common DMFT practice, prohibitively complicated generalized Pauli constraints (GPCs). Specifically, it is shown that the exact PF and EF coincide on the set of the v-representable first-order reduced density matrices (1RDMs), while FFFs approximate the common PF and EF for such 1RDMs. It is revealed that for some types of FFFs, for example for the geminal-based approximations, the GPCs are essentially trivialized to the pairwise natural occupation numbers (NONs) degeneracy conditions, routinely imposed for closed-shell systems.

Original languageEnglish
Article number012509
Pages (from-to)1-6
Number of pages6
JournalPhysical Review A
Volume100
Issue number1
DOIs
Publication statusPublished - 15 Jul 2019

Fingerprint

functionals
matrices
approximation
occupation
energy

Cite this

@article{6516e79af1b04e8b9b0418a6f7acd905,
title = "Approximating one-matrix functionals without generalized Pauli constraints",
abstract = "Common features are identified in the properties of the pure (PF) and ensemble (EF) energy functionals defined within the density matrix functional theory (DMFT) as well as their free-form (FFF) approximations. These features rationalize abandoning, in the common DMFT practice, prohibitively complicated generalized Pauli constraints (GPCs). Specifically, it is shown that the exact PF and EF coincide on the set of the v-representable first-order reduced density matrices (1RDMs), while FFFs approximate the common PF and EF for such 1RDMs. It is revealed that for some types of FFFs, for example for the geminal-based approximations, the GPCs are essentially trivialized to the pairwise natural occupation numbers (NONs) degeneracy conditions, routinely imposed for closed-shell systems.",
author = "Gritsenko, {Oleg V.} and Katarzyna Pernal",
year = "2019",
month = "7",
day = "15",
doi = "10.1103/PhysRevA.100.012509",
language = "English",
volume = "100",
pages = "1--6",
journal = "Physical Review A",
issn = "2469-9926",
publisher = "American Physical Society",
number = "1",

}

Approximating one-matrix functionals without generalized Pauli constraints. / Gritsenko, Oleg V.; Pernal, Katarzyna.

In: Physical Review A, Vol. 100, No. 1, 012509, 15.07.2019, p. 1-6.

Research output: Contribution to JournalArticleAcademicpeer-review

TY - JOUR

T1 - Approximating one-matrix functionals without generalized Pauli constraints

AU - Gritsenko, Oleg V.

AU - Pernal, Katarzyna

PY - 2019/7/15

Y1 - 2019/7/15

N2 - Common features are identified in the properties of the pure (PF) and ensemble (EF) energy functionals defined within the density matrix functional theory (DMFT) as well as their free-form (FFF) approximations. These features rationalize abandoning, in the common DMFT practice, prohibitively complicated generalized Pauli constraints (GPCs). Specifically, it is shown that the exact PF and EF coincide on the set of the v-representable first-order reduced density matrices (1RDMs), while FFFs approximate the common PF and EF for such 1RDMs. It is revealed that for some types of FFFs, for example for the geminal-based approximations, the GPCs are essentially trivialized to the pairwise natural occupation numbers (NONs) degeneracy conditions, routinely imposed for closed-shell systems.

AB - Common features are identified in the properties of the pure (PF) and ensemble (EF) energy functionals defined within the density matrix functional theory (DMFT) as well as their free-form (FFF) approximations. These features rationalize abandoning, in the common DMFT practice, prohibitively complicated generalized Pauli constraints (GPCs). Specifically, it is shown that the exact PF and EF coincide on the set of the v-representable first-order reduced density matrices (1RDMs), while FFFs approximate the common PF and EF for such 1RDMs. It is revealed that for some types of FFFs, for example for the geminal-based approximations, the GPCs are essentially trivialized to the pairwise natural occupation numbers (NONs) degeneracy conditions, routinely imposed for closed-shell systems.

UR - http://www.scopus.com/inward/record.url?scp=85069881337&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85069881337&partnerID=8YFLogxK

U2 - 10.1103/PhysRevA.100.012509

DO - 10.1103/PhysRevA.100.012509

M3 - Article

VL - 100

SP - 1

EP - 6

JO - Physical Review A

JF - Physical Review A

SN - 2469-9926

IS - 1

M1 - 012509

ER -