Some optimal pointwise ergodic theorems with rate (Q2272014)

The main result of this paper is the following Theorem 2.5. Let \(T\) be a Dunford-Schwartz operator on \((X,\Sigma,\mu)\). Let \(b\) be a slowly varying function, \(\alpha\in(0,1)\), \(p>1\), and \[ B=\sum_{n=1}^\infty \beta_nT^n, \] where \[ \beta_n={C\over n}\sum_{k\geq n}{b(k)\over k^{1+\alpha}},\quad\sum\beta_n=1. \] Let \(f\in(I-B)L^p(Z,\Sigma,\mu)\). Then \[ {\sum_{k=1}^nT^kf\over n^{1/p}(\sum_{k=1}^nb(k)^q/k^{q^\alpha})^{1/q}}\to 0\quad \mu\text{-a.s.} \] Moreover, if \((X,\Sigma,\mu)\) is non-atomic and \(T\) is induced by an ergodic invertible measure preserving \(\theta\), then for every positive function \(\psi\) satisfying \[ \lim_{n\to\infty}{(\sum_{k=1}^nb(k)^q/k^{q^\alpha})^{1/q}\over \psi(n)}=\infty, \] then there exists \(f\in(I-B)L^p(X,\Sigma,\mu)\) such that \[ \limsup_{n\to\infty}{|\sum_{k=1}^nf\circ\theta^k|\over n^{1/p}\psi(n)}=\infty\quad\mu\text{-a.s.} \] Moreover, under several conditions, one studies the convergence using Theorem 2.5 (Theorem 2.6,Theorem 2.7, and Theorem 2.8). For example, Theorem 2.8. Let \((X,\Sigma,\mu,\theta)\) be a dynamical system. Let \(b_0\) be any slowly varying function with \[ \sum_{n\geq1}{b_0(n)^2\over n}<\infty. \] Then for every \(f\in L^2(X,\Sigma,\mu)\) such that \[ \sum_{n\geq1}{||S_n(f)||_2^2\over n^2b_0(n)^2}<\infty, \] we have \[ {1\over\sqrt n}S_n(f)\to 0\quad\mu\text{-a.s.} \] Moreover, the series \(\sum_{n\geq1}{f\circ\theta^n\over\sqrt n}\) converges \(\mu\)-a.s. On the other hand, if \(\theta\) is invertible and the system is ergodic and non-atomic, for every slowly varying function \(b_1\) with \[ \sum_{n\geq1}{b_1(n)^2\over n}=\infty, \] there exists a function \(f\in L^2(X,\Sigma,\mu)\) such that \[ \sum_{n\geq1}{||S_n(f)||_2^2\over n^2b_1(n)^2}<\infty \] and \[ \limsup|{1\over\sqrt n}S_n(f)|=\infty\quad\mu\text{-a.s.} \]