Convergence to stable laws for a class of multidimensional stochastic recursions

Abstract: We consider a Markov chain

{X_{n}}_{n = 0}^{8}

on

R^{d}

defined by the stochastic recursion

X_{n} = M_{n} X_{n - 1} + Q_{n}

, where

(Q_{n}, M_{n})

are i.i.d. random variables taking values in the affine group

H = R^{d} t i m e s m G L (R^{d})

. Assume that

M_{n}

takes values in the similarity group of

R^{d}

, and the Markov chain has a unique stationary measure

u

, which has unbounded support. We denote by

| M_{n} |

the expansion coefficient of

M_{n}

and we assume

E | M |^{a} = 1

for some positive

a

. We show that the partial sums

S_{n} = s u m_{k = 0}^{n} X_{k}

, properly normalized, converge to a normal law (

a g e 2

) or to an infinitely divisible law, which is stable in a natural sense (

a < 2

). These laws are fully nondegenerate, if

u

is not supported on an affine hyperplane. Under a natural hypothesis, we prove also a local limit theorem for the sums

S_{n}

. If

a l e 2

, proofs are based on the homogeneity at infinity of

u

and on a detailed spectral analysis of a family of Fourier operators

P_{v}

considered as perturbations of the transition operator

P

of the chain

X_{n}

. The characteristic function of the limit law has a simple expression in terms of moments of

u

(

a > 2

) or of the tails of

u

and of stationary measure for an associated Markov operator (

a l e 2

). We extend the results to the situation where

M_{n}

is a random generalized similarity.

Recommendations

Cites Work

Cited In (18)

This page was built for publication: Convergence to stable laws for a class of multidimensional stochastic recursions