English

Position distribution in a generalised run and tumble process

Statistical Mechanics 2021-01-27 v1

Abstract

We study a class of stochastic processes of the type dnxdtn=v0σ(t)\frac{d^n x}{dt^n}= v_0\, \sigma(t) where n>0n>0 is a positive integer and σ(t)=±1\sigma(t)=\pm 1 represents an `active' telegraphic noise that flips from one state to the other with a constant rate γ\gamma. For n=1n=1, it reduces to the standard run and tumble process for active particles in one dimension. This process can be analytically continued to any n>0n>0 including non-integer values. We compute exactly the mean squared displacement at time tt for all n>0n>0 and show that at late times while it grows as t2n1\sim t^{2n-1} for n>1/2n>1/2, it approaches a constant for n<1/2n<1/2. In the marginal case n=1/2n=1/2, it grows very slowly with time as lnt\sim \ln t. Thus the process undergoes a {\em localisation} transition at n=1/2n=1/2. We also show that the position distribution pn(x,t)p_n(x,t) remains time-dependent even at late times for n1/2n\ge 1/2, but approaches a stationary time-independent form for n<1/2n<1/2. The tails of the position distribution at late times exhibit a large deviation form, pn(x,t)exp[γtΦn(xx(t))]p_n(x,t)\sim \exp\left[-\gamma\, t\, \Phi_n\left(\frac{x}{x^*(t)}\right)\right], where x(t)=v0tn/Γ(n+1)x^*(t)= v_0\, t^n/\Gamma(n+1). We compute the rate function Φn(z)\Phi_n(z) analytically for all n>0n>0 and also numerically using importance sampling methods, finding excellent agreement between them. For three special values n=1n=1, n=2n=2 and n=1/2n=1/2 we compute the exact cumulant generating function of the position distribution at all times tt.

Keywords

Cite

@article{arxiv.2009.01487,
  title  = {Position distribution in a generalised run and tumble process},
  author = {David S. Dean and Satya N. Majumdar and Hendrik Schawe},
  journal= {arXiv preprint arXiv:2009.01487},
  year   = {2021}
}

Comments

24 pages, 4 Figures

R2 v1 2026-06-23T18:17:11.143Z