scientific article; zbMATH DE number 3998921

From MaRDI portal
Publication:4725448





[For the entire collection see Zbl 0606.00022.] Let W be the space of media C(\({\mathbb{R}}\to [0,\infty))\cap \{W: W(0)=0\}\) and for each \(W\in {\mathbb{W}}\) let \(P_ W\) denote the probability law of the 1-dimensional diffusion process with generator \[ 2^{- 1}e^{W(x)}\frac{d}{dx}(e^{-W(x)}\frac{d}{dx}) \] and starting at 0. \(P_ W\) is a probability measure on \(\Omega =C([0,\infty)\to {\mathbb{R}})\) whose generic element is denoted by \(\omega\) with value \(\omega\) (t) at time t. Let Q be the probability measure on \({\mathbb{W}}\) such that \(\{\) W(x), \(x\geq 0\), \(Q\}\) and \(\{\) W(-x), \(x\geq 0\), \(Q\}\) are independent reflecting Brownian motions on [0,\(\infty)\) starting at 0. \(\{\omega\) (t), \(t\geq 0\}\) is regarded as a process defined on the probability space (\(\Omega\times {\mathbb{W}},P)\) where \(P(\cdot)=\int_{{\mathbb{W}}}P_ W(\cdot)Q(dW).\) The main result of this paper is that the probability distribution of (log t)\({}^{-2}\omega (t)\) converges as \(t\to \infty\) to the symmetric distribution \(\mu\) defined by \[ \int^{\infty}_{0}e^{-\lambda x}\mu (dx)=\int^{\infty}_{0}\frac{\sinh \sqrt{2\lambda}}{\sqrt{2\lambda}\cosh \sqrt{2\lambda}+t \sinh \sqrt{2\lambda}}\cdot \frac{dt}{(1+t)^ 2},\quad \lambda >0. \] For a similar result in the case of nonpositive reflected Brownian medium, see the author, Limit distributions for one-dimensional diffusion processes in self-similar random environments. to appear. Also for general discussions see \textit{K. Kawazu}, \textit{Y. Tamura} and the author, One- dimensional diffusions and random walks in random environments. Proc. 5th Japan-USSR Symp. Probab. Theory (to appear).











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4725448)