A stochastic grain growth model based on a variational principle for dissipative systems

A stochastic model for the evolution of a cellular network driven by dissipative forces is presented. The model is based on a variational formulation for the dissipated power, from which we obtain an expression for the transition-rate generating function to be used in kinetic Monte Carlo simulations. The model canonical variables are the positions and velocities of the network vertices where cell walls meet. We apply such a model to the study of grain growth in two dimensions, in which the network represents a cross-section of a polycrystalline microstructure and the cell walls represent grain boundaries. The results of the stochastic grain-growth model for relevant statistical quantities are compared to deterministic model results and analytic theories.