Scaling limit for trap models on \(\mathbb Z^d\) (Q2460329)

The work establishes scaling limit for Boushaud's model of trapped random walk on \(\mathbb Z^d\), \(d \geq 2\). This model is the nearest neighbor continuous time Markov process \(X(t)\) given by jump rates \(c(x,y)=(2d\tau_x)^{-1}\) if \(x\) and \(y\) are nearest neighbors in \(\mathbb Z^d\) and zero otherwise, where \(\{\tau_x:x \in \mathbb Z^d\}\) are i.i.d. heavy-tailed random variables distributed as \(P[\tau_x \geq u]=u^{-\alpha}(1+L(u))\) with some \(\alpha \in (0,1)\) and \(L(u) \to 0\) as \(u \to \infty\). It means that \(X(t)\) waits at a site \(x\) an exponentially distributed time with mean \(\tau_x\), and then it jumps to one of the neighbors of \(x\) with uniform probability. It is assumed that \(X(0)=0.\) Let \(B_d(t)\) be a standard \(d\)-dimensional Brownian motion started at 0, let \(V_\alpha\) be an independent \(\alpha\)-stable subordinator, and let \(Z_{d,\alpha}(s)=B_d(V^{-1}_\alpha(s))\) be the fractional-kinetic process. In the article the sequence of elements \[ X_N(t) = \frac{\sqrt{d}X(tN)}{f(N)} \] for properly chosen \(f(N)\) is considered. The main result of the work states that for a.e. random environment \(\{\tau_x: x \in \mathbb Z^d\}\) the distribution of \(X_N(t)\) converges weakly to the distribution of \(Z_{d,\alpha}\) as \(N \to \infty\). This result is in sharp contrast to the scaling limit for the same model in \(d=1\), where the limiting process is a singular diffusion in random environment.