English

How linear reinforcement affects Donsker's Theorem for empirical processes

Probability 2020-05-26 v1 Statistics Theory Statistics Theory

Abstract

A reinforcement algorithm introduced by H.A. Simon \cite{Simon} produces a sequence of uniform random variables with memory as follows. At each step, with a fixed probability p(0,1)p\in(0,1), U^n+1\hat U_{n+1} is sampled uniformly from U^1,,U^n\hat U_1, \ldots, \hat U_n, and with complementary probability 1p1-p, U^n+1\hat U_{n+1} is a new independent uniform variable. The Glivenko-Cantelli theorem remains valid for the reinforced empirical measure, but not the Donsker theorem. Specifically, we show that the sequence of empirical processes converges in law to a Brownian bridge only up to a constant factor when p<1/2p<1/2, and that a further rescaling is needed when p>1/2p>1/2 and the limit is then a bridge with exchangeable increments and discontinuous paths. This is related to earlier limit theorems for correlated Bernoulli processes, the so-called elephant random walk, and more generally step reinforced random walks.

Keywords

Cite

@article{arxiv.2005.11986,
  title  = {How linear reinforcement affects Donsker's Theorem for empirical processes},
  author = {Jean Bertoin},
  journal= {arXiv preprint arXiv:2005.11986},
  year   = {2020}
}
R2 v1 2026-06-23T15:47:03.670Z