2019
DOI: 10.1103/physreve.100.023307
|View full text |Cite
|
Sign up to set email alerts
|

Fermion sign problem in path integral Monte Carlo simulations: Quantum dots, ultracold atoms, and warm dense matter

Abstract: The ab initio thermodynamic simulation of correlated Fermi systems is of central importance for many applications, such as warm dense matter, electrons in quantum dots, and ultracold atoms. Unfortunately, path integral Monte Carlo (PIMC) simulations of fermions are severely restricted by the notorious fermion sign problem (FSP). In this work, we present a hands-on discussion of the FSP and investigate in detail its manifestation with respect to temperature, system size, interactionstrength and -type, and the d… Show more

Help me understand this report
View preprint versions

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

9
168
0

Year Published

2019
2019
2022
2022

Publication Types

Select...
8

Relationship

3
5

Authors

Journals

citations
Cited by 123 publications
(177 citation statements)
references
References 112 publications
9
168
0
Order By: Relevance
“…ii) the current validity domain of, e.g., the GDSMFBB parametrization to 0 ≤ r s ≤ 20 can be significantly extended by incorporating the recent ab initio PIMC results for the electron liquid regime (20 ≤ r s ≤ 100) by Dornheim et al [128]. iii) new ab intio QMC results at low temperature, θ ∼ 10 −1 , could help to more accurately resolve open thermodynamic questions like the effective mass enhancement [129], but are difficult to obtain due to the notorious fermion sign problem [130]. iv) Neural networks are known to be valuable as universal function approximators (see also Sec.…”
Section: A Summary Of Ab Initio Static Resultsmentioning
confidence: 99%
“…ii) the current validity domain of, e.g., the GDSMFBB parametrization to 0 ≤ r s ≤ 20 can be significantly extended by incorporating the recent ab initio PIMC results for the electron liquid regime (20 ≤ r s ≤ 100) by Dornheim et al [128]. iii) new ab intio QMC results at low temperature, θ ∼ 10 −1 , could help to more accurately resolve open thermodynamic questions like the effective mass enhancement [129], but are difficult to obtain due to the notorious fermion sign problem [130]. iv) Neural networks are known to be valuable as universal function approximators (see also Sec.…”
Section: A Summary Of Ab Initio Static Resultsmentioning
confidence: 99%
“…At θ = 0.5, the simulations are computationally expensive due to the fermion sign problem (we find an average sign of S ≈ 0.01, see Ref. [104] for an extensive discussion) and the statistical uncertainty is substantial for large q. Still, the neural net is capable to provide a prediction that is fully consistent with the benchmark data even in a regime where the training data are sparse.…”
Section: B Physical Behavior and Neural Net Representationmentioning
confidence: 91%
“…1. Here the main obstacle is given by the notorious fermion sign problem [104,105], which prevents PIMC simulations below half the Fermi temperature and limits the feasible system size to N = 14, . .…”
Section: Arxiv:190708473v1 [Physicsplasm-ph] 19 Jul 2019mentioning
confidence: 99%
See 1 more Smart Citation
“…Moreover, it is straightforward to see that the sign exponentially decreases both with β and the system size N , which is the origin of the notorious fermion sign problem [104,105], see Ref. [106] for a topical review article. More specifically, the relative statistical uncertainty from a fermionic Monte Carlo calculation is inversely proportional to S [107],…”
Section: Path Integral Monte Carlomentioning
confidence: 99%