scientific article; zbMATH DE number 139883

From MaRDI portal
Publication:4028223





[For the entire collection see Zbl 0747.00029.] Let \(m\) denote Lebesgue measure on \([0,1]\). A linear operator \(T\) on \(L^ 1(m)\) is called stochastic if \(T\) takes probability densities to probability densities. We denote by \({\mathcal S}\) the set of all stochastic operators and we write \({\mathcal S}_{\leq\lambda}=\{T\in{\mathcal S}\): \(T\lambda\leq\lambda\}\), \({\mathcal S}_ \lambda=\{T\in{\mathcal S}\): \(T\lambda=\lambda\}\) for any equivalent sigma-finite measure \(\lambda\). If \(T\in S_{\leq\lambda}\), we let \(Pf={dm \over d\lambda} T({d\lambda \over dm}f)\) and we say that \(T\) is mixing if \[ \lim \int_ B P^ n 1_ A d\lambda=\lambda(A)\lambda(B)/\lambda(0,1) \] whenever \(\lambda(A)+\lambda(B)<\infty\) [see e.g. \textit{S. R. Foguel}, The ergodic theory of Markov processes (1969; Zbl 0282.60037)]. We investigate the size of the set of mixing operators. It is shown that this set is residual in \({\mathcal S}\leq \lambda\) with respect to both the operator norm and the strong operator topologies. On the other hand it is meager for the weak operator topology. The same holds true in \({\mathcal S}_ \lambda\). The residuality result has been extended to the stronger notion of completely mixing in the more recent joint paper of the author and \textit{R. Rȩbowski} [Ann. Pol. Math. 56, No. 3, 233-242 (1992)].











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 Q4028223)