English

Convergence to stable laws for a class of multidimensional stochastic recursions

Probability 2008-11-10 v2

Abstract

We consider a Markov chain {Xn}n=0\8\{X_n\}_{n=0}^\8 on Rd\R^d defined by the stochastic recursion Xn=MnXn1+QnX_{n}=M_n X_{n-1}+Q_n, where (Qn,Mn)(Q_n,M_n) are i.i.d. random variables taking values in the affine group H=RdGL(Rd)H=\R^d\rtimes {\rm GL}(\R^d). Assume that MnM_n takes values in the similarity group of Rd\R^d, and the Markov chain has a unique stationary measure ν\nu, which has unbounded support. We denote by Mn|M_n| the expansion coefficient of MnM_n and we assume \EM\a=1\E |M|^\a=1 for some positive \a\a. We show that the partial sums Sn=k=0nXkS_n=\sum_{k=0}^n X_k, properly normalized, converge to a normal law (\a2\a\ge 2) or to an infinitely divisible law, which is stable in a natural sense (\a<2\a<2). These laws are fully nondegenerate, if ν\nu is not supported on an affine hyperplane. Under a natural hypothesis, we prove also a local limit theorem for the sums SnS_n. If \a2\a\le 2, proofs are based on the homogeneity at infinity of ν\nu and on a detailed spectral analysis of a family of Fourier operators PvP_v considered as perturbations of the transition operator PP of the chain {Xn}\{X_n \}. The characteristic function of the limit law has a simple expression in terms of moments of ν\nu (\a>2\a > 2) or of the tails of ν\nu and of stationary measure for an associated Markov operator (\a2\a\le 2). We extend the results to the situation where MnM_n 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}
}
R2 v1 2026-06-21T11:24:02.889Z