Convergence to stable laws for a class of multidimensional stochastic recursions
Abstract
We consider a Markov chain on defined by the stochastic recursion , where are i.i.d. random variables taking values in the affine group . Assume that takes values in the similarity group of , and the Markov chain has a unique stationary measure , which has unbounded support. We denote by the expansion coefficient of and we assume for some positive . We show that the partial sums , properly normalized, converge to a normal law () or to an infinitely divisible law, which is stable in a natural sense (). These laws are fully nondegenerate, if is not supported on an affine hyperplane. Under a natural hypothesis, we prove also a local limit theorem for the sums . If , proofs are based on the homogeneity at infinity of and on a detailed spectral analysis of a family of Fourier operators considered as perturbations of the transition operator of the chain . The characteristic function of the limit law has a simple expression in terms of moments of () or of the tails of and of stationary measure for an associated Markov operator (). We extend the results to the situation where is a random generalized similarity.
Keywords
Cite
@article{arxiv.0809.4349,
title = {Convergence to stable laws for a class of multidimensional stochastic recursions},
author = {Dariusz Buraczewski and Ewa Damek and Yves Guivarc'h},
journal= {arXiv preprint arXiv:0809.4349},
year = {2008}
}