Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

arXiv:cond-mat/0110245 · The Journal of Chemical Physics

An adaptive, kink-based approach to path integral calculations

Randall W. Hall

Published Jan 1, 2002• 16 citations
arXivHTMLPDF

Abstract

A kink-based expression for the canonical partition function is developed using Feynman’s path integral formulation of quantum mechanics and a discrete basis set. The approach is exact for a complete set of states. The method is tested on the 3×3 Hubbard model and overcomes the sign problem seen in traditional path integral studies of fermion systems. Kinks correspond to transitions between different N-electron states, much in the same manner as occurs in configuration interaction calculations in standard ab initio methods. The different N-electron states are updated, based on which states occur frequently during a Monte Carlo simulation, giving better estimates of the true eigenstates of the Hamiltonian.

DOI
10.1063/1.1423939

Related papers

Path integral studies of the 2D Hubbard model using a new projection operator

Randall W. Hall

Feynman’s path integral formulation of quantum mechanics, supplemented by an approximate projection operator (exact in the case of noninteracting particles), is used to study the 2D Hubbard model. The projection operator is designed to study Hamiltonians defined on a finite basis set, but extensions to continuous basis sets are suggested. The projection operator is shown to reduce the variance by a significant amount relative to straightforward Monte Carlo integration. Approximate calculations are usually within one standard deviation of exact results and virtually always within two to three standard deviations. In addition, the algorithm scales with the number of discretization points P as either P or P2 (depending on the method of implementation), rather than the P3 of the Hubbard–Stratonovich transformation. Accuracy to about 5%–10% in energies and spin–spin correlation functions are found using moderate amounts of computer time.

The Journal of Chemical Physics•Jan 15, 1991•5 citations

Sign Learning Kink-based (SiLK) Quantum Monte Carlo for molecular systems

Xiaoyao Ma, Randall W. Hall, Frank Löffler +4 more

The Sign Learning Kink (SiLK) based Quantum Monte Carlo (QMC) method is used to calculate the ab initio ground state energies for multiple geometries of the H2O, N2, and F2 molecules. The method is based on Feynman's path integral formulation of quantum mechanics and has two stages. The first stage is called the learning stage and reduces the well-known QMC minus sign problem by optimizing the linear combinations of Slater determinants which are used in the second stage, a conventional QMC simulation. The method is tested using different vector spaces and compared to the results of other quantum chemical methods and to exact diagonalization. Our findings demonstrate that the SiLK method is accurate and reduces or eliminates the minus sign problem.

arXiv:1512.05455•The Journal of Chemical Physics•Jan 5, 2016•0 citations

Multilevel Blocking Approach to the Fermion Sign Problem in Path-Integral Monte Carlo Simulations

Chi H. Mak, Reinhold Egger, H. Weber-Gottschick

A general algorithm toward the solution of the fermion sign problem in finite-temperature quantum Monte Carlo simulations has been formulated for discretized fermion path integrals with nearest-neighbor interactions in the Trotter direction. This multilevel approach systematically implements a simple blocking strategy in a recursive manner to synthesize the sign cancellations among different fermionic paths throughout the whole configuration space. The practical usefulness of the method is demonstrated for interacting electrons in a quantum dot.

arXiv:cond-mat/9810002•Physical Review Letters•Nov 23, 1998•70 citations

Improved methods for path integral Monte Carlo integration in fermionic systems

William H. Newman, Atsuo Kuki

We generalize the discretized Feynman path integral expansion by replacing the path through pure states with a path through idempotent density matrices. The transformed expression converges to the ordinary path integral, but is computationally more flexible than the ordinary form. By introducing a particular choice of these idempotent density matrices, based on rotational averaging around the two-particle center of mass, we greatly reduce the sign problem for systems of two fermions in three dimensions. In the ordinary path integral Monte Carlo approach, low temperature simulation of fermions is inefficient as the integral decays exponentially faster than the integrand (and its variance) as the temperature decreases. The new rotationally averaged algorithm dramatically retards this relative decay. Fermionic simulations of the model system of Kestner and Sinanoglu, and of the harmonic oscillator, demonstrate this improvement, as integrals calculated at temperatures much lower than the singlet–triplet splitting display negligible decay of the desired averages relative to the partition function of the modified path ensemble.

The Journal of Chemical Physics•Jan 15, 1992•16 citations

Calculation of Canonical Properties of a Quantum System by Path Integral Numerical Methods

Mikhail Voznesenskiy, P. N. Vorontsov‐Velyaminov

Abstract We propose a new approach in obtaining the partition function of a system of several interacting particles (fermions) in external field within path integral Monte Carlo method based on direct averaging of the exchange contributions over the positive weight determined by non‐closed trajectories. This procedure is performed at fixed temperatures, ranging to rather low values. The complete partition function for each temperature is then obtained within an expanded ensemble procedure. We found a good agreement of data obtained by the new method with results of previously proposed path‐integral‐expanded‐ensemble Monte Carlo calculations [1, 2]. The new approach allows to reach observably lower temperatures compared to method [1, 2], thus providing a significant reduce of the sign problem. It yields an independent way to treat thermal properties of quantum systems, so the good agreement with previous data [1, 2] allows us to test and state the validity of both approaches. Simulations for systems of 2 up to 5 non‐interacting particles and particles interacting with Coulomb repulsion in 3D harmonic field were performed. The presented approach allows one to carry out calculations for low temperatures that makes it possible to extract data for the ground‐state energy and low‐temperature thermodynamics. (© 2013 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)

Contributions to Plasma Physics•May 1, 2013•0 citations

Functional integral approach: A third formulation of quantum statistical mechanics

Xian Xi Dai, William Edwin Evenson

Quantum statistical mechanics has developed primarily through two approaches, pioneered by Gibbs and Feynman, respectively. In Gibbs' method one calculates partition functions from phase-space integrations or sums over stationary states. Alternatively, in Feynman's approach, the focus is on the path-integral formulation. The Hubbard-Stratonovich transformation leads to a functional-integral formulation for calculating partition functions. We outline here the functional integral approach to quantum statistical mechanics, including generalizations and improvements to Hubbard's formulation. We show how the dimensionality of the integrals is reduced exactly, how the problem of assuming an unknown canonical transformation is avoided, how the reality of the partition function in the complex representation is guaranteed, and how the extremum conditions are simplified. This formulation can be applied to general systems, including superconductors.

Physical review. E, Statistical physics, plasmas, fluids, and related interdisciplinary topics•Jan 18, 2002•5 citations
A kink-based expression for the canonical partition function is developed using Feynman’s path integral formulation of quantum mechanics and a discrete basis set. The approach is exact for a complete set of states. The method is tested on the 3×3 Hubbard model and overcomes the sign problem seen in traditional path integral studies of fermion systems. Kinks correspond to transitions between different N-electron states, much in the same manner as occurs in configuration interaction calculations in standard ab initio methods. The different N-electron states are updated, based on which states occur frequently during a Monte Carlo simulation, giving better estimates of the true eigenstates of the Hamiltonian.
Feynman’s path integral formulation of quantum mechanics, supplemented by an approximate projection operator (exact in the case of noninteracting particles), is used to study the 2D Hubbard model. The projection operator is designed to study Hamiltonians defined on a finite basis set, but extensions to continuous basis sets are suggested. The projection operator is shown to reduce the variance by a significant amount relative to straightforward Monte Carlo integration. Approximate calculations are usually within one standard deviation of exact results and virtually always within two to three standard deviations. In addition, the algorithm scales with the number of discretization points P as either P or P2 (depending on the method of implementation), rather than the P3 of the Hubbard–Stratonovich transformation. Accuracy to about 5%–10% in energies and spin–spin correlation functions are found using moderate amounts of computer time.
The Sign Learning Kink (SiLK) based Quantum Monte Carlo (QMC) method is used to calculate the ab initio ground state energies for multiple geometries of the H2O, N2, and F2 molecules. The method is based on Feynman's path integral formulation of quantum mechanics and has two stages. The first stage is called the learning stage and reduces the well-known QMC minus sign problem by optimizing the linear combinations of Slater determinants which are used in the second stage, a conventional QMC simulation. The method is tested using different vector spaces and compared to the results of other quantum chemical methods and to exact diagonalization. Our findings demonstrate that the SiLK method is accurate and reduces or eliminates the minus sign problem.
A general algorithm toward the solution of the fermion sign problem in finite-temperature quantum Monte Carlo simulations has been formulated for discretized fermion path integrals with nearest-neighbor interactions in the Trotter direction. This multilevel approach systematically implements a simple blocking strategy in a recursive manner to synthesize the sign cancellations among different fermionic paths throughout the whole configuration space. The practical usefulness of the method is demonstrated for interacting electrons in a quantum dot.
We generalize the discretized Feynman path integral expansion by replacing the path through pure states with a path through idempotent density matrices. The transformed expression converges to the ordinary path integral, but is computationally more flexible than the ordinary form. By introducing a particular choice of these idempotent density matrices, based on rotational averaging around the two-particle center of mass, we greatly reduce the sign problem for systems of two fermions in three dimensions. In the ordinary path integral Monte Carlo approach, low temperature simulation of fermions is inefficient as the integral decays exponentially faster than the integrand (and its variance) as the temperature decreases. The new rotationally averaged algorithm dramatically retards this relative decay. Fermionic simulations of the model system of Kestner and Sinanoglu, and of the harmonic oscillator, demonstrate this improvement, as integrals calculated at temperatures much lower than the singlet–triplet splitting display negligible decay of the desired averages relative to the partition function of the modified path ensemble.
Abstract We propose a new approach in obtaining the partition function of a system of several interacting particles (fermions) in external field within path integral Monte Carlo method based on direct averaging of the exchange contributions over the positive weight determined by non‐closed trajectories. This procedure is performed at fixed temperatures, ranging to rather low values. The complete partition function for each temperature is then obtained within an expanded ensemble procedure. We found a good agreement of data obtained by the new method with results of previously proposed path‐integral‐expanded‐ensemble Monte Carlo calculations [1, 2]. The new approach allows to reach observably lower temperatures compared to method [1, 2], thus providing a significant reduce of the sign problem. It yields an independent way to treat thermal properties of quantum systems, so the good agreement with previous data [1, 2] allows us to test and state the validity of both approaches. Simulations for systems of 2 up to 5 non‐interacting particles and particles interacting with Coulomb repulsion in 3D harmonic field were performed. The presented approach allows one to carry out calculations for low temperatures that makes it possible to extract data for the ground‐state energy and low‐temperature thermodynamics. (© 2013 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim)
Quantum statistical mechanics has developed primarily through two approaches, pioneered by Gibbs and Feynman, respectively. In Gibbs' method one calculates partition functions from phase-space integrations or sums over stationary states. Alternatively, in Feynman's approach, the focus is on the path-integral formulation. The Hubbard-Stratonovich transformation leads to a functional-integral formulation for calculating partition functions. We outline here the functional integral approach to quantum statistical mechanics, including generalizations and improvements to Hubbard's formulation. We show how the dimensionality of the integrals is reduced exactly, how the problem of assuming an unknown canonical transformation is avoided, how the reality of the partition function in the complex representation is guaranteed, and how the extremum conditions are simplified. This formulation can be applied to general systems, including superconductors.