We propose a new method for quantum Monte Carlo simulations with the negative-sign problem. In the conventional method, one simulates a virtual system whose Boltzmann factors are the absolute values of the original one, and analyzes the data so that the correct expression for physical quantities may be recovered. We simulate a different virtual system with properties closer to the original system and yet without the negative-sign problem. Comparison between the present method and the conventional one is made for the antiferromagnetic Heisenberg model on the 4×h4 square lattice with frustrations ( J 1 - J 2 model). We conclude that the present method is useful in the low-temperature region where the negative-sign problem becomes serious. It is also shown that the method is considerably efficient on the point that quantities for different parameters can be obtained in a single run.
Quantum simulations are a powerful tool for exploring strongly correlated many-body phenomena. Yet, their reach is limited by the fermion sign problem, which causes configuration weights to become negative, compromising statistical sampling. In auxiliary-field Quantum Monte Carlo calculations of the doped Hubbard model, neglecting the sign
S of the weight leads to qualitatively wrong results -- most notably, an apparent suppression rather than enhancement of
d-wave pairing at low temperature. Here we approach the problem from a different perspective: instead of identifying negative-weight paths, we examine the statistics of measured observables in a sign-resolved manner. By analyzing histograms of key quantities (kinetic energy, antiferromagnetic structure factor, and pair susceptibilities) for configurations with
S=±1, we derive an exact relation linking the bias from ignoring the sign to the difference between sign-resolved means,
Δμ, and the average sign,
⟨S⟩. Our framework provides a precise diagnostic of the origin of measurement bias in Quantum Monte Carlo and clarifies why observables such as the
d-wave susceptibility are especially sensitive to the sign problem.
Sign problem in quantum Monte Carlo (QMC) simulation appears to be an extremely hard yet interesting problem. In this article, we present a pedagogical overview on the origin of the sign problem in various quantum Monte Carlo simulation techniques, ranging from the world-line and stochastic series expansion Monte Carlo for boson and spin systems to the determinant and momentum-space quantum Monte Carlo for interacting fermions. We point out the basis dependency of the sign problem and summarize the progresses to cure, ease and even make use of the sign problem over the years, such as symmetry analysis of the underlying Hamiltonian, basis optimization in writting down the partition functions and many others. Moreover, we state that although traditional lore saying that in case of sign problem, the average sign in QMC simulation approaches zero exponentially fast with the space-time volume of the configurational space, there are recent breakthroughs showing this is not always the case and based on the properties of the partition function for finite size systems, one could distinguish when the average sign has the usual exponential scaling and when it is bestowed with an algebraic scaling at the low temperature limit. Fermionic QMC simulations with such algebraic sign problems have been successfully carried out for extended Hubbard-type and quantum Moiré lattice models.