Minimally entangled typical thermal state algorithms for finite temperature Matsubara Green functions
Jul 29, 2021
11 pages
Published in:
- Phys.Rev.B 105 (2022) 19, 195107,
- Phys.Rev.B 105 (2022) 195107
- Published: May 6, 2022
e-Print:
- 2107.13941 [cond-mat.str-el]
DOI:
- 10.1103/PhysRevB.105.195107 (publication)
View in:
Citations per year
Abstract: (APS)
We extend finite temperature tensor network methods to compute Matsubara imaginary time correlation functions, building on the minimally entangled typical thermal states (METTS) and purification algorithms. While imaginary time correlation functions are straightforward to formulate with these methods, care is needed to avoid convergence issues that would result from naive estimators. As a benchmark, we study the single-band Anderson impurity model, even though the algorithm is quite general and applies to lattice models. The special structure of the impurity model benchmark system and our choice of basis enable techniques such as the reuse of high-probability METTS for increasing algorithm efficiency. The results are competitive with state-of-the-art continuous-time Monte Carlo. We discuss the behavior of computation time and error as a function of the number of purified sites in the Hamiltonian.Note:
- 12 pages, 11 figures
- model: lattice
- correlation function
- renormalization group
- thermal
- finite temperature
- density matrix
- network
- Hamiltonian
- site
- efficiency
References(45)
Figures(11)
- [1]
- [2]
- [3]
- [4]
- [5]
- [6]
- [7]
- [8]
- [9]
- [10]
- [11]
- [12]
- [13]
- [14]
- [15]
- [16]
- [17]
- [18]
- [19]
- [20]
- [21]
- [22]
- [23]
- [24]
- [25]