Berry–Esseen bounds and moderate deviations for random walks on GLd(R)

Autor: Ion Grama, Quansheng Liu, Hui Xiao
Přispěvatelé: Universitat Hildesheim, Institut fur Mathematik and Angewandte Informatik, Hildesheim, Germany, Laboratoire de Mathématiques de Bretagne Atlantique (LMBA), Université de Bretagne Sud (UBS)-Université de Brest (UBO)-Centre National de la Recherche Scientifique (CNRS)
Rok vydání: 2021
Předmět:
Zdroj: Stochastic Processes and their Applications
Stochastic Processes and their Applications, Elsevier, 2021, 142, pp.293-318. ⟨10.1016/j.spa.2021.08.005⟩
ISSN: 0304-4149
DOI: 10.1016/j.spa.2021.08.005
Popis: Let ( g n ) n ⩾ 1 be a sequence of independent and identically distributed random elements of the general linear group G L d ( R ) , with law μ . Consider the random walk G n : = g n … g 1 . Denote respectively by ‖ G n ‖ and ρ ( G n ) the operator norm and the spectral radius of G n . For log ‖ G n ‖ and log ρ ( G n ) , we prove moderate deviation principles under exponential moment and strong irreducibility conditions on μ ; we also establish moderate deviation expansions in the normal range [ 0 , o ( n 1 / 6 ) ] and Berry–Esseen bounds under the additional proximality condition on μ . Similar results are found for the couples ( X n x , log ‖ G n ‖ ) and ( X n x , log ρ ( G n ) ) with target functions, where X n x : = G n ⋅ x is a Markov chain and x is a starting point on the projective space P ( R d ) .
Databáze: OpenAIRE