Calculations of the distant-pair model show that the Monte Carlo and approximate analytic results do not always agree. Dunstan's comparison of the calculated photoluminescence decays is shown to be an insufficient test of his method. The comparison of the metastable population densities is shown to be invalid.
We apply our Monte Carlo twist method to the antiferromagnetic Ising model on square lattices in several uniform fields imposing fixed boundary conditions. We have estimated the critical temperature T c and critical exponent ν at each field. The former well agree with the analytic results of Müller-Hartmann and Zittartz, and the latter are all close to ν=1. We have calculated the finite-size critical-point amplitude of the interfacial free energy Ξ 0 which is believed to be universal. Our estimate of Ξ 0 at zero field is in good agreement with the analytic result of Abraham and Švrakić. We also report for the first time estimates of Ξ 0 at finite fields. Further, comparison is made with results of Mon and Jasnow.
We present a quantum Monte Carlo method which allows calculations on many-fermion systems at finite temperatures without any sign decay. This enables simulations of the grand-canonical ensemble at large system sizes and low temperatures. Both diagonal and off-diagonal expectations can be computed straightforwardly. The sign decay is eliminated by a constraint on the fermion determinant. The algorithm is approximate. Tests on the Hubbard model show that accurate results on the energy and correlation functions can be obtained.