New multiple histogram method for studying phase transitions

Abstract

The density of states contains the complete information about a thermodynamical system. We present a multiple histogram method for the purpose of efficiently calculating the density of states from the data of a computer simulation (e.g. a Monte Carlo simulation). Using the dynamical ensemble, our method shows no critical slowing down in the vicinity of a first order phase transition. Furthermore for most applications it is not necessary to join the histograms. As a first application we determine the order of the phase transition in the modified xy model in dependence on the interaction parameter.