Fractional Brownian motion with mean-density interaction

Abstract: Fractional Brownian motion is a Gaussian stochastic process with long-range correlations in time; it has been shown to be a useful model of anomalous diffusion. Here, we investigate the effects of mutual interactions in an ensemble of particles undergoing fractional Brownian motion. Specifically, we introduce a mean-density interaction in which each particle in the ensemble is coupled to the gradient of the total, time-integrated density produced by the entire ensemble. We report the results of extensive computer simulations for the mean-square displacements and the probability densities of particles undergoing one-dimensional fractional Brownian motion with such a mean-density interaction. We find two qualitatively different regimes, depending on the anomalous diffusion exponent $\alpha$ characterizing the fractional Gaussian noise. The motion is governed by the interactions for $\alpha < 4/3$ whereas it is dominated by the fractional Gaussian noise for $\alpha > 4/3$. We develop a scaling theory explaining our findings. We also discuss generalizations to higher space dimensions and nonlinear interactions as well as applications to the growth of strongly stochastic axons (e.g., serotonergic fibers) in vertebrate brains.