The connectedness of Sierpiński sponges with rotational and reflectional components and associated graph-directed systems
Huo-Jun Ruan, Jian-Ci Xiao
TL;DR
The paper addresses the connectedness of sponge-like Cantor-type graph-directed attractors in low- and high-dimensional spaces when rotations and reflections are allowed. It reduces the problem to a coordinate-wise intersection test using auxiliary graphs $G_P$ for faces of the unit cube and a global intersection graph with solid and dashed edges, showing that $K_i ∩ K_j$ is nonempty iff there exists an infinite solid walk or a terminated finite walk starting from $(i,j)$. A finite-iteration criterion is established with a computable depth $c(n,d)$ so that $K_i ∩ K_j ≠ ∅$ is detected by level-$c(n,d)$ approximations $Q_{i,c(n,d)}$, enabling practical connectivity checks via Hata graphs of symmetry orbits. The authors further relate sponge connectivity to Hata-graphs of the symmetry family $(O(F))_{O∈ ext{O}_d}$, and provide improvements on iteration bounds (e.g., $F_{d+1}$ suffices when all rotations are identity) with extensions to broader graph-directed systems. Overall, the work furnishes actionable, finite criteria for testing connectedness of complex sponge-like fractals with rotational/reflectional components and generalizes to broader graph-directed frameworks."
Abstract
We provide two methods to characterize the connectedness of all $d$-dimensional generalized Sierpiński sponges whose corresponding IFSs are allowed to have rotational and reflectional components. Our approach is to reduce it to an intersection problem between the coordinates of graph-directed attractors. More precisely, let $(K_1,\ldots,K_n)$ be a Cantor-type graph-directed attractor in $\mathbb{R}^d$. By creating an auxiliary graph, we provide an effective criterion for whether $K_i\cap K_j$ is empty for every pair of $1\leq i,j\leq n$. Moreover, the emptiness can be checked by examining only a finite number of geometric approximations of the attractor. The approach is also applicable to more general graph-directed systems.
