Embedding loose trees in $k$-uniform hypergraphs
Yaobin Chen, Allan Lo
TL;DR
This work advances the theory of spanning tree embeddings to loose trees in $k$-uniform hypergraphs by proving that for every $k\ge 4$, an $n$-vertex hypergraph with minimum $(k-2)$-degree at least $(1/2+\gamma)\binom{n}{k-2}$ contains every spanning loose tree of bounded degree, for large $n$. The authors develop a robust framework based on absorption, weak hypergraph regularity, and a robust reduced-graph structure that supports fractional matchings, reachability, and rotatability. A key innovation is handling the $(k,k-2)$-threshold regime where a tight Hamilton cycle in the reduced graph cannot be guaranteed, achieved via a careful construction of a robust spanning subgraph $G^*$ and a sequence of embedding steps (absorption, almost spanning embedding, and final absorption). The results establish that the loose-tree embedding threshold matches the corresponding perfect-matching threshold for $$k\ge 4$$, thereby resolving a central case and extending prior works on the $k=3$ setting. The findings have significant implications for spanning structures in hypergraphs and illuminate a robust embedding paradigm applicable to broader hypergraph embedding problems.
Abstract
A classical result of Komlós, Sárközy and Szemerédi shows that every large $n$-vertex graph with minimum degree at least $(1/2+γ)n$ contains all spanning trees of bounded degree. We generalised this result to loose spanning hypertrees in $k$-uniform hypergraphs, that is, linear hypergraphs obtained by subsequently adding edges sharing a single vertex with a previous edge. We give a general sufficient condition for embedding loose trees with bounded degree. In particular, we show that for all $k\ge 4$, every $n$-vertex $k$-uniform hypergraph with $n\ge n_0(k,γ, Δ)$ and minimum $(k-2)$-degree at least $(1/2+γ)\binom{n}{k-2}$ contains every spanning loose tree with maximum vertex degree at most $Δ$. This bound is asymptotically tight. This generalises a result of Pehova and Petrova, who proved the case when $k=3$ and of Pavez-Signé, Sanhueza-Matamala and Stein, who considered the codegree threshold for bounded degree tight trees.
