paper

Highest Trees of Random Mappings

arXiv:1504.04532

Abstract

We study the heights of the trees of a random mapping of elements. Using singularity analysis of exact generating functions, we prove that the mapping has a unique highest tree with probability . The property of having a unique highest tree plays a crucial role in the solution of the Road Coloring Problem [Trahtman, 2009]. More generally, we consider -branches, the subtrees rooted at distance from the cycles. For all fixed and we show that the highest -branch exceeds every other -branch in height by at least , except with probability asymptotic to . We also show that, for any fixed , with probability the part of the highest -branch that lies above the second highest one (its crown) has more than times as many vertices as roots. The last result is used in the author's proof that a random -letter automaton with states is synchronizing with probability .

References in corpus (1)

Cited by in corpus (1)