Monotone runs of uniformly distributed integer random variables: a probabilistic analysis

  • Authors:
  • Guy Louchard

  • Affiliations:
  • Département d'Informatique, Université Libre de Bruxelles, Boulevard du Triomphe, Brusselles, Belgium

  • Venue:
  • Theoretical Computer Science - In memoriam: Alberto Del Lungo (1965-2003)
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Using a Markov chain approach and a polyomino-like description, we study some asymptotic properties of monotone increasing runs of uniformly distributed integer random variables. We analyze the limiting trajectories, which after suitable normalization, lead to a Brownian motion, the number of runs, which is asymptotically Gaussian, the run length distribution, the hitting time to a large length k run, which is asymptotically exponential, and the maximum run length which is related to the Gumbel extreme-value distribution function. A preliminary application to DNA analysis is also given.