English

Block size in Geometric(p)-biased permutations

Probability 2018-08-15 v4

Abstract

Fix a probability distribution p=(p1,p2,)\mathbf p = (p_1, p_2, \cdots) on the positive integers. The first block in a p\mathbf p-biased permutation can be visualized in terms of raindrops that land at each positive integer jj with probability pjp_j. It is the first point KK so that all sites in [1,K][1,K] are wet and all sites in (K,)(K,\infty) are dry. For the geometric distribution pj=p(1p)j1p_j= p(1-p)^{j-1} we show that plogKp \log K converges in probability to an explicit constant as pp tends to 0. Additionally, we prove that if p\mathbf p has a stretch exponential distribution, then KK is infinite with positive probability.

Cite

@article{arxiv.1708.05626,
  title  = {Block size in Geometric(p)-biased permutations},
  author = {Irina Cristali and Vinit Ranjan and Jake Steinberg and Erin Beckman and Rick Durrett and Matthew Junge and James Nolen},
  journal= {arXiv preprint arXiv:1708.05626},
  year   = {2018}
}

Comments

10 pages, new title

R2 v1 2026-06-22T21:18:00.878Z