Brownian survival among Poissonian traps with random shapes at critical intensity (Q1779991)

The authors consider a standard Brownian motion \(\{B(s):s\geq 0\}\) in \( \mathbb{R}^{d},~d\geq 2.\) Let \(\mathcal{C}\) be the family of all closed subsets of \(\mathbb{R}^{d},\) assume that \(\mathcal{C}\) is endowed with the topology generated by the Hausdorff metric, and let \(\Pi \) be a probability measure on the Borel subsets of \(\mathcal{C}.\) Let \(\{\omega _{t}:t\geq 0\}\) be a Poisson point process with intensity \(\nu _{t}=ct^{-1}\log ^{2}t\) if \( d=2,\) and \(=ct^{-2/d}\) if \(d\geq 3,\) where \(c\in (0,\infty ).\) For \(t\geq 0,\) let \(K_{t}=\bigcup_{x\in \omega _{t}}(x+A_{x}),\) where \(A_{x},~x\in \omega _{t},\) are i.i.d. random sets drawn from \(\mathcal{C},\) and write \(\tau _{K_{t}}=\inf \{s\geq 0:B(s)\in K_{t}\}.\) Then \(K_{t}\) is viewed as a collection of randomly located and randomly shaped traps, and \(\tau _{K_{t}}\) is viewed as the trapping time for the Brownian motion. For \(t\geq 0,\) let \( S_{t}\) denote the probability of survival up to time \(t.\) Under appropriate conditions on \(\Pi ,\) it is first shown that \[ S_{t}=E\left[ \exp \left( -\nu _{t}\int_{\mathcal{Q}}\Pi (dA)\left| W^{A}(t)\right| \right) \right] , \] where \(\mathcal{Q}\) is the family of all compact subsets \(A\) of \(\mathbb{R }^{d}\) with \(\emptyset \neq A=(A^{\circ })^{-},\) and \(\{W^{A}(t):t\geq 0\}\) is the Wiener sausage with shape \(A\in \mathcal{Q}.\) The goal of this paper is to identify the asymptotic behaviour of \(S_{y}\) for large \(t.\) The main results show that the tail of \(S_{t}\) has an interesting dependence on \(c,\) with one regime for \(d=2,\) and two regimes depending on \(\Pi \) for \(d\geq 3.\) The proofs rely heavily on an earlier paper of the authors [Ann. Math. (2) 153, 355--406 (2001; Zbl 1004.60021)].