Ageing in the parabolic Anderson model (Q1944665)

One of the most important subjects of disordered complex systems out of equilibrium is their long term dynamics. A key paradigm in this research is the notion of ageing. This process was observed in many different systems and the main aim of this paper is to investigate whether ageing can also be observed for a different kind of dynamics, given by a diffusion or heat flow with an inhomogeneous potential. Such a dynamics describes, for example, the intensity of populations in inhomogeneous environments or a chemical reactant in the presence of an inhomogeneously distributed catalytic substance. However, there is some controversy in the literature around the question whether such systems exhibit ageing. In some publications, a special case is used, namely, a parabolic Anderson model with time-variable potential to show absence of correlation-based ageing. But, in this paper, the authors show a presence of ageing (in the original sense) for the parabolic Anderson model if the underlying random potential is sufficiently heavy-tailed. This work presents three main results. The first one, Theorem 1.1, shows that the probability that, during the time window \([t,t+\theta t]\), the profiles of the solution of the parabolic Anderson problem remain within distance \(\epsilon >0\) of each other converges to a constant \(I (\theta)\), which is strictly between zero and one. This shows that ageing holds on a linear time scale. The second main result, Theorem 1.3, is an almost sure ageing result. It defines a function \(R(t)\) which characterizes the waiting time starting from time \(t\) until the profile changes again. The third main result, Theorem 1.6, is a functional scaling limit theorem for the location of the peak, which determines the profile, and for the growth rate of the solution.