Critical dynamics of finite Ising model

The Glauber kinetics of Ising spins is considered as a queueing process and simulated event by event as first proposed by Bortz, Kalos, and Lebowitz. The advantage of this algorithm compared to the standard single-flip Monte Carlo method is discussed for the situation of slowing down of dynamics. This process is used to generate fluctuations of magnetization and energy in the critical regimeT=T_c of two-dimensional Ising models. The analysis of these fluctuations leads to numerical determination of the critical exponents for dynamics: for the size dependence of correlation time atT _c , and for frequency dependence of the power spectrumS()~ ^–µ . From the finite-size scaling hypothesis, scaling relations are settled which are confirmed by this numerical experiment.