WorksMathematics
Local Limit Theorems and Edgeworth Expansions for some classes of inhomogeneous Markov chains
Abstract
In this note, we prove local central limit theorems and expansions in the central limit theorem for inhomogeneous Markov chains satisfying a one-step lower ellipticity condition. As for the local limit theorem, in the general inhomogeneous case, compared with \cite{DS}, we are able to prove results without two-sided ellipticity, although only in the one-step case. Compared with \cite{PeligradLLT1,PeligradLLT2}, we are able to prove local limit theorems without a lower -mixing condition. Concerning Edgeworth expansions, compared with \cite{DHPTRF1}, we are able to get results without two-sided ellipticity (under some sequential Diophantine conditions). Our conditions are then verified for Markov chains in a random dynamical environment with a one-sided Doeblin condition under very mild assumptions. For Markov chains in a random environment, we also consider the more general case of -mixing chains and then, under some mild periodicity assumptions on the environment, we prove local central limit theorems and first-order Edgeworth expansions (without any ellipticity assumptions), matching the results for deterministic homogeneous chains. When the chain is driven by a topologically mixing subshift of finite type, we obtain high-order expansions for -mixing chains under some Diophantine conditions of a deterministic system generated by a periodic point. We then apply our results to bounded additive functionals of posterior states in disordered quantum trajectories and establish local central limit theorems and first-order Edgeworth expansions.