Matrix Product States with Large Sites
Abstract
We explore various ways to group orbitals into clusters in a matrix product state (MPS). We explain how a generic cluster MPS can often lead to an increase in computational cost and instead propose a special cluster structure, involving only the first and last orbitals/sites, with a wider scope for computational advantage. This structure is a natural formalism to describe correlated multireference (MR) theories. We demonstrate the flexibility and usefulness of this approach by implementing various uncontracted MR configuration interaction, perturbation, and linearized coupled cluster theories using an MPS with large cluster sites. Applications to the nitrogen dimer, the chromium dimer, and benzene, including up to triple excitations in the external space, demonstrate the utility of an MPS with up to two large sites. We use our results to analyze the quality of different multireference approximations.
Additional Information
© 2022 American Chemical Society. Received 22 September 2021. Published online 21 January 2022. We thank Cyrus Umrigar for providing us with the SHCI PEC data for the chromium dimer. Work by G.K.C. was supported by the US National Science Foundation (NSF) via Grant CHE-2102505. G.K.C. acknowledges additional support from the Simons Foundation via the Many-Electron Collaboration and the Investigator Award. Work by H.Z. was supported by the US Department of Energy, Office of Science, Basic Energy Sciences under Award DE-SC0019374. Work by H.R.L. was supported by the Air Force Office of Scientific Research, under Award FA9550-18-1-0095. H.R.L. acknowledges support from a postdoctoral fellowship from the German Research Foundation (DFG) via Grant LA 4442/1-1 during the first part of this work. K.G. acknowledges support from a Ph.D. fellowship and a travel grant for a long stay abroad at the California Institute of Technology from the Research Foundation Flanders (FWO Vlaanderen). Some of the computations presented here were conducted at the Resnick High Performance Computing Center, a facility supported by the Resnick Sustainability Institute at the California Institute of Technology. The authors declare the following competing financial interest(s): G.K.C. is a part owner of QSimulate, Inc.Attached Files
Accepted Version - 2109.11036.pdf
Supplemental Material - ct1c00957_si_001.pdf
Files
Name | Size | Download all |
---|---|---|
md5:7be4816fa74d200d7b35ba46c9fdf7fc
|
1.0 MB | Preview Download |
md5:fd8eae469e0ae78cb638a0fb48056dd6
|
282.0 kB | Preview Download |
Additional details
- Eprint ID
- 113094
- DOI
- 10.1021/acs.jctc.1c00957
- Resolver ID
- CaltechAUTHORS:20220125-893621641
- NSF
- CHE-2102505
- Simons Foundation
- Department of Energy (DOE)
- DE-SC0019374
- Air Force Office of Scientific Research (AFOSR)
- FA9550-18-1-0095
- Deutsche Forschungsgemeinschaft (DFG)
- LA 4442/1-1
- Fonds Wetenschappelijk Onderzoek (FWO)
- Created
-
2022-01-25Created from EPrint's datestamp field
- Updated
-
2022-04-06Created from EPrint's last_modified field
- Caltech groups
- Resnick Sustainability Institute