Bereiche | Tage | Auswahl | Suche | Downloads | Hilfe

DY: Dynamik und Statistische Physik

DY 46: Poster

DY 46.92: Poster

Donnerstag, 27. März 2003, 15:30–18:00, P1

Dynamical properties of the Worm Algorithm — •Christine Gabriel and Hans Gerd Everttz — Institut für Theoretische Physik, Petersgasse 16, 8010 Graz, Austria

We study the dynamics of the “Worm Algorithm”, a novel approach to efficient Monte Carlo simulations of many classical statistical systems [PRL, 87(16), 2001]. The “Worm Algorithm” utilizes the idea of updating closed path configurations of the high-temperature representation through the motion of end points of a disconnected path. We have investigated this method for the case of the Ising model and compare it to cluster algorithms. We find that the integrated autocorrelation time of the energy grows only logarithmically. For spatial correlations and for the susceptibility we find very small autocorrelation times, even on the largest lattices. We have extended the algorithm to include external fields via the “ghost-spin” formalism, where we also do not see any large autocorrelations even in a strong magnetic field, where simulations with other methods are difficult. We have applied the extended “Worm Algorithm” to simulations of “capillary condensation”.