No label defined (Q4940261)

Let \(X\) be a set and let \({\mathcal B}\) be a \(\sigma\)-algebra of subsets of \(X.\) The author defines a random measure \(\mu\) as a set of random variables (not only nonnegative) \(\{\mu(A),A\in {\mathcal B} \}\) such that \(\mu(A\cup B)=\mu(A)+\mu(B)\) a.s. for \(A\cap B=\emptyset\) and \(\mu(A_{n}) @>P>>0\) as \(A_{n}\downarrow\emptyset.\) For a random variable \(\xi\) the author considers a seminorm \(\|\xi\|=\sup\{\delta\colon P\{|\xi|> \delta\}>\delta\}.\) The set of \(\mu\)-integrable functions \(f_{\alpha}: X\to R\), \(\alpha\in U\), is uniformly integrable on \(\mu\) if for all \(\varepsilon>0\) \[ \lim_{C\to +\infty} \sup_{\alpha,A}P\Biggl\{ \Biggl|\int_{A\cap\{|f_{\alpha} |>C\}}f_{\alpha} d\mu\Biggr|>\varepsilon\Biggr\}=0. \] The author proves the following assertion: A set of functions \(f_{\alpha}: X\to R\), \(\alpha\in U\), which are integrable with respect to a random measure \(\mu\), is uniformly integrable on \(\mu\) if and only if there exists a measurable function \(G(t): R\to R\) such that \(\lim_{|t|\to\infty}|G(t)/t|=+\infty\) and the set of random variables \(\{\int_{A} G(f_{\alpha}(x)) d\mu\), \(A\in {\mathcal B}\), \(\alpha\in U\}\) is bounded in probability.