Pointwise convergence of nonconventional averages

Measure-preserving transformations (28D05) Dynamical aspects of measure-preserving transformations (37A05) Ergodic theorems, spectral theory, Markov operators (37A30) Ergodic theory of linear operators (47A35)

Review text

Let \((X,\mathcal B, \mu)\) be a finite Lebesgue measure space. Consider operators \(U_i\), \(i=1,2\), on \(L^p(\mu)\), \(1<p<\infty\), positive together with their inverses, and such that \(\text{sup}_{n\in \mathbb Z}\| U_i^n\| _p<\infty\). The first part of the paper is an answer to the following question by H. Furstenberg: Do we have pointwise convergence of the averages \[ \frac{1}{N}\sum_{n=1}^N U_1^n[f\cdot U_2^ng](x) \] for all bounded functions \(f\) and \(g\)? It is shown that there is no pointwise, or even weak convergence if \(U_2\) is a negative power of \(U_1\). However, when \(U_2\) is a positive power of \(U_1\), the above averages converge a.e. for functions in \(L^p\). It is also shown that for a positive contraction \(V\) in \(L^p\), \(1<p<\infty\), the averages \[ \frac{1}{N}\sum_{n=1}^N V^n[f\cdot V^ng](x) \] converge almost everywhere. Now, consider two dynamical systems \((X,\mathcal B, \mu,T_i)\), \(i=1,2\), where \(T_i:X\to X\) are measure-preserving transformations. The second part of the paper deals with the convergence of \[ \frac{1}{N}\sum_{n=1}^N f(T_1^nx)g(T_2^nx). \] The author gives several sufficient conditions for the averages above to converge a.e. for all functions \(f\) and \(g\) in \(L^2\).

Recommendations

Cited in

(13)

This page was built for publication: Pointwise convergence of nonconventional averages

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4677437)