On the invariant measure of a positive recurrent diffusion in R

DOI10.1007/S10959-006-0046-XMaRDI QIDQ877232zbMATH OpenOpenAlexFDO

Publication date 19 April 2007

Published in Journal of Theoretical Probability (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/math/0412410

diffeomorphism diffusion invariant measure stochastic flow

Dynamical systems and their relations with probability theory and stochastic processes (37A50) Stochastic partial differential equations (aspects of stochastic analysis) (60H15) Diffusion processes (60J60)

Abstract: Given an one-dimensional positive recurrent diffusion governed by the Stratonovich SDE [ X_t=x+int_0^tsigma(X_s)strat db(s)+int_0^t m(X_s) ds, ] we show that the associated stochastic flow of diffeomorphisms focuses as fast as

m a t h r m e x p (- 2 t i n t_{R} f r a c m^{2} s i g m a^{2} d P i)

, where

d P i

is the finite stationary measure. Moreover, if the drift is reversed and the diffeomorphism is inverted, then the path function so produced tends, independently of its starting point, to a single (random) point whose distribution is

d P i

. Applications to stationary solutions of

X_{t}

, asymptotic behavior of solutions of SPDEs and random attractors are offered.

Review text

Let \(X_t\) be the positive recurrent diffusion in \(\mathbb{R}\) solving the Stratonovich equation \[ X_t= x+ \int^t_0 \sigma(X_s)\circ db_s+ \int^t_0 m(X_s)\,ds. \] Then for all \(x,y\in\mathbb{R}\) almost surely \[ \lim_{t\to\infty}(1/t)\ln|s(X_t(x))- s(X_t(y))|= -2\int_{\mathbb{R}} (m^2/\sigma^2)\,d\Pi, \] where \(s(.)\) is the scale function and \(\Pi\) the invariant measure of \(X_t\). The expression on the right also provides an upper bound for the spectral gap of \(X_t\). Let \(X^\#_t\) be the diffusion solving \[ X^\#_t= x+ \int^t_0\sigma(X^\#_s)\circ db_s- \int^t_0 m(X^\#_s)\,ds, \] and \(X^\#_t(x,{\mathfrak b})\) the corresponding flow. Then there exists a random variable \({\mathcal X}\), distributed according to \(\Pi\), such that almost surely \(X^\#_t(x,{\mathfrak b})_{t\to\infty}\), whenever \(x>{\mathcal X}\), and \(X^\#_t(x,{\mathfrak b})_{t\to\infty}\), if \(x<{\mathcal X}\). It is shown that almost surely \(\liminf_{t\to\infty}|X^\#_t({\mathcal X})|<\infty\), furthermore \({\mathbf P}(\lim_{t\to\infty}|X^\#_t|=\infty\) for every \(x\neq{\mathcal X})\) and \({\mathbf P}(\lim_{t\to\infty}X^\#_t(x)= {\mathcal X}\) for all \(x)= 1\), where \(X^\#_t\) is constructed as follows: For any \(T> 0\) let \(b_T(t):= b(T- t)- b(T)\), \(0\leq t\leq T\), let \(\{X^T_t\), \(0\leq t\leq T\}\) be the solution of \(d^T_t= \sigma(X^T_t)\circ db_T(t)+ m(X^T_t)\,dt\), \(X^T_0= x\), and define \(X^\downarrow_T:= X^T_T\). As a simple example, \(X_t\) is taken to be the Ornstein-Uhlenbeck process. Three applications are given: \({\mathcal X}\) can be used to obtain stationary solutions for \(X_t\) and \(X^\#_t\). For every smooth function \(f\), \(f(X^\downarrow_t)\) uniquely solves \[ f(X^\#_t)= f(x)- \sigma(x)\int^t_0(\partial f(X^\downarrow_t)/\partial x)\,db(s)+ \int^t_0{\mathcal G} f(X^\downarrow_s)\,ds, \] where \({\mathcal G}\) is the infinitesimal generator of \(X_t\) and \({\mathcal X}\) can be identified as the random attractor of the corresponding random dynamical system.

Recommendations

Cites work

Cited in

(6)

This page was built for publication: On the invariant measure of a positive recurrent diffusion in \({\mathbb{R}}\)

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q877232)