Abstract:The path integral formulation of quantum mechanical problems including fermions is often affected by a severe numerical sign problem. We show how such a sign problem can be alleviated by a judiciously chosen constant imaginary offset to the path integral. Such integration contour deformations introduce no additional computational cost to the Hybrid Monte Carlo algorithm, while its effective sample size is greatly increased. This makes otherwise unviable simulations efficient for a wide range of parameters. Applying our method to the Hubbard model, we find that the sign problem is significantly reduced. Furthermore, we prove that it vanishes completely for large chemical potentials, a regime where the sign problem is expected to be particularly severe without imaginary offsets. In addition to a numerical analysis of such optimized contour shifts, we analytically compute the shifts corresponding to the leading and next-to-leading order corrections to the action. We find that such simple approximations, free of significant computational cost, suffice in many cases.
Strongly Correlated Electrons,High Energy Physics - Lattice,Computational Physics
What problem does this paper attempt to address?
The problem that this paper attempts to solve is the Fermionic Sign Problem in the path integral of quantum mechanics. In quantum systems containing fermions, the path - integral method is often severely affected by the numerical sign problem, which leads to inefficiency or impossibility of numerical simulations. Specifically, when the Hamiltonian of the system contains terms that cause complex phases, the sign problem in the path integral is particularly severe, making it extremely difficult to calculate the partition function and its related physical quantities.
### Main contributions of the paper
1. **Proposed a solution**:
- By introducing a carefully selected constant imaginary - part shift to deform the integration contour of the path integral, the authors alleviated the sign problem. This method does not require additional computational cost and can significantly increase the effective sample size, making simulations that were originally difficult to achieve efficient and feasible.
2. **Theoretical analysis**:
- The authors proved that for the case of large chemical potential, the sign problem completely disappears, which is the parameter region where the sign problem is usually most severe.
- The authors also obtained the offsets corresponding to the leading and sub - leading corrections of the action through analytical calculations, and these simple approximations are already effective enough in many cases.
3. **Numerical verification**:
- The authors applied this method to the Hubbard model and found that the sign problem was significantly reduced.
- Through numerical analysis and optimization, the authors further verified the effectiveness of this method and showed how to maximize the statistical efficiency by optimizing the constant offset.
### Specific methods
1. **Path deformation of path integral**:
- By deforming the integration domain from the real axis to a certain favorable manifold in the complex plane, sign oscillations can be reduced. This deformation is allowed without crossing the singularities of the integrand and while maintaining the integral homology class.
- In particular, the authors studied the simplest deformation method, that is, deforming the integration manifold by a global constant imaginary - part shift.
2. **Saddle points and quantum corrections**:
- The authors determined the optimal constant offset by solving the saddle - point equation. Saddle points are points where the complex phase is fixed in the path integral and can be found by analytical methods or numerical optimization.
- The authors also considered the influence of quantum corrections on the position of saddle points, and these corrections correspond to the contributions of one - particle irreducible diagrams.
3. **Numerical optimization algorithm**:
- To further improve the statistical efficiency, the authors proposed a search algorithm based on the Newton - Raphson method, predicting the position of the optimal offset by calculating the first - and second - order derivatives of the statistical efficiency with respect to the imaginary - part offset.
- This algorithm can find an offset that is better than a simple offset or a saddle - point offset in most cases, thereby significantly reducing the required number of samples and computational resources.
### Conclusion
This paper successfully alleviated the Fermionic Sign Problem by introducing the method of constant imaginary - part offset, especially in the case of large chemical potential. This method has been strictly proven theoretically and verified through numerical experiments. In addition, the proposed optimization algorithm provides an effective tool for further improving computational efficiency. These results are of great significance for the study of quantum many - body systems, especially for the study of strongly correlated electron systems and QCD at finite baryon density.