Using Monte Carlo methods, we can simulate various phenomena in classical many-body systems. But Monte Carlo simulations for quantum systems are suffered from some particular problems for a long time. Recent developments of quantum Monte Carlo algorithms provide significant improvement in Monte Carlo simulations for some important quantum models. In this paper, such recent developments are reviewed from a unified perspective. In particular, two algorithms-loop and directed loop algorithms-are presented. Firstly, we will give the basis of quantum Monte Carlo method with path-integral representation. Quantum Monte Carlo algorithm gives importance sampling of configurations on path-integral representation, but a conventional one is not efficient. Secondly, we will introduce new novel variables in path-integral representation, which called "graph variable". And we will present loop and directed loop algorithms using such new graph variables.
抄録全体を表示