A note on the scaling limits of random Pólya trees
Author(s) -
Bernhard Gittenberger,
Emma Yu Jin,
Michael Wallner
Publication year - 2017
Language(s) - English
Resource type - Conference proceedings
DOI - 10.1137/1.9781611974775.8
Subject(s) - mathematics , combinatorics , infinity , tree (set theory) , limit (mathematics) , scaling , random tree , scaling limit , random graph , discrete mathematics , mathematical analysis , geometry , graph , robot , motion planning , computer science , artificial intelligence
Panagiotou and Stufler (arXiv:1502.07180v2) recently proved one important fact on their way to establish the scaling limits of random Pu0027{o}lya trees: a uniform random Pu0027{o}lya tree of size $n$ consists of a conditioned critical Galton-Watson tree $C_n$ and many small forests, where with probability tending to one as $n$ tends to infinity, any forest $F_n(v)$, that is attached to a node $v$ in $C_n$, is maximally of size $vert F_n(v)vert=O(log n)$. Their proof used the framework of a Boltzmann sampler and deviation inequalities. In this paper, first, we employ a unified framework in analytic combinatorics to prove this fact with additional improvements on the bound of $vert F_n(v)vert$, namely $vert F_n(v)vert=Theta(log n)$. Second, we give a combinatorial interpretation of the rational weights of these forests and the defining substitution process in terms of automorphisms associated to a given Pu0027{o}lya tree. Finally, we derive the limit probability that for a random node $v$ the attached forest $F_n(v)$ is of a given size.
Accelerating Research
Robert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom
Address
John Eccles HouseRobert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom