Multiple ergodic averages for flows and an application (Q351768)

Let \(\{T_t\}_{t\in\mathbb{R}}\) be a measure preserving flow on a Lebesque space~\({(X, \mu)}\) and \({p_1,\ldots,p_k:\mathbb{R}^d}\to\mathbb{R}\) any collection of polynomials. Let \({\Pi_d\subset\mathbb{R}^d}\) be a rectangle \({[0,R_1]\times\ldots\times[0,R_d]}\) and \({|\Pi_d|=R_1\dots R_d}.\) The main result of the article under review states that for any \({k\in\mathbb{N}}\) and \({f_1,\ldots, f_k\in L^\infty(X,\mu)}\) \[ \frac{1}{|\Pi_d|}\int_{\Pi_d}f_1\circ T_{p_1(s)}\dots f_k\circ T_{p_k(s)}\,ds \] converges in \({L^2(X,\mu)}\) as \({R_1,\ldots, R_d\to\infty}.\) If the polynomials are \(\mathbb{R}\)-independent then the limit equals to the product of integrals \[ \int_Xf_1\,d\mu \,\dots\int_Xf_k\,d\mu. \] The proof of convergence is based on the technique of \textit{B. Host} and \textit{B. Kra} is seminorms from [Ann. Math. (2) 161, No. 1, 397--488 (2005; Zbl 1077.37002)], some induction argument as in [\textit{V. Bergelson}, Ergodic Theory Dyn. Syst. 7, 337--349 (1987; Zbl 0645.28012)] and is reduced to the case of nilflows. The proof of identification of the limit is based on methods given in [\textit{N. Frantzikinakis} and \textit{B. Kra}, Isr. J. Math. 148, 267--276 (2005; Zbl 1155.37303)] and [\textit{A.~Leibman} Trans. Am. Math. Soc. 362, No. 3, 1619--1658 (2010; Zbl 1191.37016)]. The formula for the limit is also given when the polynomials are not necessarily independent. The case when each \({f_j=1_A}\) for some measurable set \(A\) is considered.