Dynamic critical exponent of some Monte Carlo algorithms for the self-avoiding walk

Abstract

Discusses the dynamic critical behavior of some Monte Carlo algorithms for the self-avoiding walk (SAW). For algorithms with local N-conserving elementary moves, it is argued that the autocorrelation time behaves as tau approximately Np with p approximately=2+2 nu . For the BFACF dynamics (a grand canonical algorithm), Monte Carlo data is presented indicating that p=2.2+or-0.5 for two-dimensional non-reversal random walks and p=3.0+or-0.4 for two-dimensional SAW, values which are significantly less than 2+2 nu .