Rational Gluing in Edge Replacement Systems
Davide Perego, Matteo Tarocchi
TL;DR
The paper addresses the rationality of the gluing relation arising from edge replacement systems and constructs a finite-state automaton $\mathrm{Gl}_{\mathcal{R}}$ that recognises pairs of infinite words that are glued in the limit space. By encoding pairs of edges across expansion levels with transitions of types 0, 1, and 2, the automaton captures when two words are equivalent under the gluing relation $\sim$ on the Cantor-type symbol space $\Omega_{\mathcal{R}}$ and the limit space $X_{\mathcal{R}} = \Omega_{\mathcal{R}} / \sim$. The main result proves that the gluing relation is rational: $(x_0x_1\ldots, y_0y_1\ldots)$ is in $\sim$ if and only if the run of $(x_0,y_0)(x_1,y_1)\ldots$ in $\mathrm{Gl}_{\mathcal{R}}$ is accepting, with a careful delineation of how Type 0–Type 2 transitions reflect edge adjacency across expansion levels. The framework yields explicit automata for classical examples such as interval, Ważewski dendrite, and Basilica replacement systems, providing a computable finite-state model for the gluing relation and enabling further analysis of limit spaces and their rearrangement groups in totally disconnected settings.
Abstract
In this paper, we prove the rationality of the gluing relation of edge replacement systems, which were introduced for studying rearrangement groups of fractals. More precisely, we describe an algorithmic procedure for building a finite state automaton that recognizes pairs of equivalent sequences that are glued in the fractal. This fits in recent interest towards the rationality of gluing relations on totally disconnected compact metrizable spaces.
