A dominated ergodic theorem for some bilinear averages (Q973976)

Let \((X,{\mathcal F},\mu)\) be a \(\sigma\)-finite measure space, \(T\) be a positive invertible linear operator with positive inverse on some \(L^p (\mu)\), \(1\leq p<\infty\). The paper considers the averages \({\mathcal A}_n f(x)=\frac{1}{2n+1}\sum_{i=-n}^{n} T^i f(x)\) and the associated ergodic maximal operator \({\mathcal M}f(x)=\sup_{n\geq 0} {\mathcal A}_n |f|(x)\). The main result of the paper reads as follows. Let \(1\leq p\), \(1<p_1,p_2 <\infty\), \(\frac{1}{p_1 }+\frac{1}{p_2 }=\frac{1}{p}\). Then the following statements are equivalent: {\parindent=7mm \begin{itemize}\item[(a)] The bilinear averages \({\mathcal A}_n (f_1 ,f_2 )\) converge in \(L^p (\mu )\) as \(n\rightarrow\infty\) for all \(f_1 \in L^{p_1 }(\mu )\), \(f_2 \in L^{p_2}(\mu )\). \item[(b)] There exists \(C>0\) such that \(\sup_{n\geq 0} \|{\mathcal A}_n (f_1 ,f_2 )\|_p \leq C \| f_1 \|_{p_1 } \| f_2 \|_{p_2}\) for all \(f_1 \in L^{p_1}(\mu)\), \(f_2 \in L^{p_2}(\mu)\). \item[(c)] There exists \(C>0\) such that \(\|{\mathcal M}(f_1 ,f_2 )\|_p \leq C\| f_1 \|_{p_1} \| f_2 \|_{p_2 }\) for all \(f_1 \in L^{p_1 }(\mu )\), \(f_2 \in L^{p_2 }(\mu )\). \item[(d)] \({\mathcal A}_n (f_1 ,f_2 )\) converges a.e.\ as \(n\rightarrow\infty\) for all \(f_1 \in L^{p_1}(\mu)\), \(f_2 \in L^{p_2 }(\mu )\).\end{itemize}}