A short proof of the central limit theorem for Markov chains (Q1097584): Difference between revisions

Let \(X_ k\) be a Markov process on a probability space (X,\({\mathcal F},\mu)\). The author uses his results for semigroups of linear operators [Stochastic Processes Appl. 19, 183-187 (1985; Zbl 0568.60007)] to show that the limiting distribution of \(\sum^{n}_{k=1}\alpha (X_ k)/\sqrt{n}\) is normal. The conditions on the function \(\alpha\) and the transition probabilities are given in terms of the resolvent of a bounded linear operator but the author shows that, for a finite state space \({\mathcal X}\), these conditions are satisfied whenever \(\sum_{y\in {\mathcal X}}\pi (y)\alpha (y)=0\), where \(\pi\) (y) is the stationary distribution.