RUS  ENG
Full version
JOURNALS // Teoriya Veroyatnostei i ee Primeneniya // Archive

Teor. Veroyatnost. i Primenen., 1972 Volume 17, Issue 2, Pages 380–383 (Mi tvp2605)

This article is cited in 1 paper

Short Communications

A generalization of an ergodic theorem of Hopf

A. A. Tempel'man

Institute of Physics and Mathematics, Academy of Sciences, Lithuanian SSR

Abstract: Let $X$ be a separable locally compact semigroup; let($\Omega$, $\mathfrak G$, $m$) be a space with a $\sigma$-finite measure $m$ and let $T_x$, $x\in X$, be a dynamic system in $\Omega$ with “time” from $X$. Let, further, $p$ and $q$ be probability Borel measures on $X$ and $\lambda_n=\sum_{k=0}^np*q^{*k}$. If $f$, $g\in L_1(m)$ and $g>0$ then the limit
$$ \lim_{n\to\infty}\int_Xf(T_x\omega)\lambda_n(dx)\bigg/\int_Xg(T_x\omega)\lambda_n(dx)=h_{f,g}(\omega) $$
is shown to exist almost everywhere on $\Omega$.
$(p,q)$-conservative dynamic systems are defined as systems inducing recurrent random walks in $\Omega$ in correspondence with the measures $p$ and $q$. For such dynamic systems the equality $h_{f,g}=\mathbf E(f\mid\mathfrak F)$ is proved where $\mathbf E(f\mid\mathfrak F)$ is the conditional expectation of the function $f(\omega)$ given the $\sigma$-algebra $\mathfrak F$ of measurable invariant sets.

Received: 21.07.1970


 English version:
Theory of Probability and its Applications, 1973, 17:2, 363–365

Bibliographic databases:


© Steklov Math. Inst. of RAS, 2024