A regeneration proof of the central limit theorem for uniformly ergodic Markov chains (Q5900353)

Let \(X=(X_{n})_{n\geq 1}\) be a Markov process with measurable state space \(\left( E,\mathcal{E}\right) \) and unique stationary distribution \(\pi \) on \(\mathcal{E}\). Consider an \(\mathcal{E}\)-measurable function \(h:E\rightarrow \mathbb{R}\) such that \(\int_{E}h^{2}d\pi <\infty \). It is a classical result that if \(X\) is uniformly ergodic, then the central limit theorem (CLT) holds, that is, \(n^{-1/2}\sum_{i=1}^{n}[h(X_{i})- \int_{E}hd\pi ]\) is asymptotically normal as \(n\rightarrow \infty \). The present authors provide a regeneration proof of this result. Reviewer's remark: The CLT considered has been first proved by \textit{I. A. Ibragimov} [Dokl. Akad. Nauk SSSR 125, 711--714 (1959; Zbl 0087.13303); Theor. Probab. Appl. 7, 349--382 (1962); translation from Teor. Veroyatn. Primen. 7, 361--392 (1962; Zbl 0119.14204)]. In the more general context of weakly dependent strictly stationary processes while the authors credit R. Gogburg with such a result (published in 1972). Curiously enough, this is explicit in the 1971 monograph by \textit{I.A. Ibragimov} and \textit{Yu.V. Linnik} [Independent and stationary sequences of random variables. (1971; Zbl 0219.60027)] quoted by the present authors.