Projective convergence of columns for inhomogeneous products of matrices with nonnegative entries

preprint2014arXivOpen access

Abstract

Let PnP_n be the nn-step right product A1⋯AnA_1\cdots A_n, where A1,A2,…A_1,A_2,\dots is a given infinite sequence of d×dd\times d matrices with nonnegative entries. In a wide range of situations, the normalized matrix product Pn/∥Pn∥P_n/{\Vert P_n\Vert} does not converge and we shall be rather interested in the asymptotic behavior of the normalized columns PnUi/∥PnUi∥P_nU_i/\Vert P_nU_i\Vert, where U1,…,UdU_1,\dots,U_d are the canonical d×1d\times 1 vectors. Our main result in Theorem~A gives a sufficient condition (C){\bf (C)} over the sequence A1,A2,…A_1,A_2,\dots ensuring the existence of {\it dominant columns} of PnP_n, having the same projective limit VV: more precisely, for any rank nn, there exists a partition of {1,…,d}\{1,\dots,d\} made of two subsets Jn≠∅J_n\ne\emptyset and JncJ_n^c such that each one of the sequences of normalized columns, say PnUjn/∥PnUjn∥P_nU_{j_n}/\Vert P_nU_{j_n}\Vert with jn∈Jnj_n\in J_n tends to VV as nn tends to +∞+\infty and are {\it dominant} in the sense that the ratio ∥PnUjn′/∥PnUjn∥\Vert P_nU_{j_n'}/\Vert P_nU_{j_n}\Vert tends to 00, as soon as jn′∈Jncj_n'\in J_n^c. The existence of sequences of such {\it dominant columns} implies that for any probability vector XX with positive entries, the probability vector PnX/∥PnX∥P_nX/\Vert P_nX\Vert, converges as nn tends to +∞+\infty. Our main application of Theorem~A (and our initial motivation) is related to an {\it Erd\H os problem} concerned with a family of probability measures μβμ_β (for 1<β<21<β<2 a real parameter) fully supported by a subinterval of the real line, known as {\it Bernoulli convolutions}.

Explore this paper’s authors, topics and related work.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

Reviews 0

Write a reviewWrite

No reviews yet.

Discussion 0

Add a commentWrite

No comments yet.