Spinal partitions and invariance under re-rooting of continuum random trees

DOI10.1214/08-AOP434MaRDI QIDQ838002zbMATH OpenOpenAlexFDO

Authors Bénédicte Haas, Matthias Winkel, Jim Pitman

Publication date 21 August 2009

Published in The Annals of Probability (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/0705.3602

continuum random tree discrete tree fragmentation process Markov branching model Poisson-Dirichlet distribution random re-rooting spinal decomposition

Mathematics Subject Classification ID

Branching processes (Galton-Watson, birth-and-death, etc.) (60J80)

Abstract: We develop some theory of spinal decompositions of discrete and continuous fragmentation trees. Specifically, we consider a coarse and a fine spinal integer partition derived from spinal tree decompositions. We prove that for a two-parameter Poisson--Dirichlet family of continuous fragmentation trees, including the stable trees of Duquesne and Le Gall, the fine partition is obtained from the coarse one by shattering each of its parts independently, according to the same law. As a second application of spinal decompositions, we prove that among the continuous fragmentation trees, stable trees are the only ones whose distribution is invariant under uniform re-rooting.

Review text

The authors develop some theory of spinal decomposition of fragmentation trees that arise as genealogical trees of fragmentation processes. The focus is on Markovian partition-valued fragmentation processes in discrete time with partitions of \([n]= \{1,\dots,n\}\) and in continuous time with partitions of \(\mathbb{N}\). In the first case it is postulated that each non-singleton block splits at each time, which leads to Markov branching models. In the second case a self-similarity condition is imposed which leads to self-similar random trees. Starting from a rooted combinatorial tree \(T_{[n]}\) with \(n\) leaves labelled by \([n]\), the path from the root to the leave labelled 1 is the spine of \(T_{[n]}\). Deleting each edge along this spine defines a graph whose connected components are called bushes. If, as well as cutting every edge on the spine, each edge connected to a spinal vertex is cut, each bush is further decomposed into sub-trees. Thus two nested partitions of \(\{2,\dots, n\}\) are obtained, which naturally extend to partitions of \([n]\) by adding the singleton \(\{1\}\). These partitions of \([n]\) are called the coarse spinal partition and the fine spinal partition derived from \(T_{[n]}\). It is proved that for a two-parameter Poisson-Dirichlet family of continuous fragmentation trees, including the stable trees of \textit{T. Duquesne} and \textit{J. F. Le Gall} [Probab. Theory Relat. Fields 131, No. 4, 553--603 (2005; Zbl 1070.60076)] and generalizing the family of stable fragmentations studied by \textit{G. Miermont} [Probab. Theory Relat. Fields 127, No. 3, 423--454 (2003; Zbl 1042.60043)], that the fine partition is obtained from the coarse one by shattering each of its parts independently, according to the same law. As another application of spinal decompositions, it is shown that among the continuous fragmentation trees, stable trees are the only ones whose distribution is invariant under uniform re-rooting.

Recommendations

Cites work

Cited in

(26)

This page was built for publication: Spinal partitions and invariance under re-rooting of continuum random trees

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q838002)