PreprintarXiv licencearXiv
The threshold for fractional clique decompositions of random hypergraphs, via matrix scaling
Abstract
Let δ*k,r be the fractional Kr(k)-decomposition threshold in minimum codegree. For fixed k≥2, r≥ k+1 and ε>0, we prove that every n-vertex k-uniform hypergraph G with minimum codegree at least (δ*k,r+ε)n admits, with high probability, a fractional Kr(k)-decomposition after retaining each edge independently with probability p ≥ C(log n/nr−k)1/(binomrk−1). The bound on p is sharp up to a constant factor, confirming the conjectured threshold order in the graph case. The proof develops a probabilistic matrix-scaling approach.
Related
- Fractional clique decompositions in random hypergraphs
- Clique-factors in randomly perturbed hypergraphs
- On the distinct maximal-clique sizes in k-uniform hypergraphs
- Asymptotic Enumeration of F-decompositions in Typical Hypergraphs and of Universal Cycles
- Fractional Triangle Decompositions at Partite Minimum Degree 4n/5
BibTeX
@misc{tran2026threshold,
title = {{The threshold for fractional clique decompositions of random hypergraphs, via matrix scaling}},
author = {Tuan Tran},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.00924},
eprint = {2610.00924},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.00924},
}
RIS
TY - UNPB
TI - The threshold for fractional clique decompositions of random hypergraphs, via matrix scaling
AU - Tuan Tran
PY - 2026
DA - 2026/10/01
PB - arXiv
DO - 10.48550/arxiv.2610.00924
UR - https://arxiv.org/abs/2610.00924
AB - Let δ^*_k,r be the fractional K_r^(k)-decomposition threshold in minimum codegree. For fixed k≥2, r≥ k+1 and ε>0, we prove that every n-vertex k-uniform hypergraph G with minimum codegree at least (δ^*_k,r+ε)n admits, with high probability, a fractional K_r^(k)-decomposition after retaining each edge independently with probability p ≥ C(log n/n^r−k)^1/(binomrk−1). The bound on p is sharp up to a constant factor, confirming the conjectured threshold order in the graph case. The proof develops a probabilistic matrix-scaling approach.
ER -