English

The Minimization of Random Hypergraphs

Discrete Mathematics 2020-11-03 v3 Data Structures and Algorithms Combinatorics Probability

Abstract

We investigate the maximum-entropy model Bn,m,p\mathcal{B}_{n,m,p} for random nn-vertex, mm-edge multi-hypergraphs with expected edge size pnpn. We show that the expected size of the minimization of Bn,m,p\mathcal{B}_{n,m,p}, i.e., the number of its inclusion-wise minimal edges, undergoes a phase transition with respect to mm. If mm is at most 1/(1p)(1p)n1/(1-p)^{(1-p)n}, then the minimization is of size Θ(m)\Theta(m). Beyond that point, for α\alpha such that m=1/(1p)αnm = 1/(1-p)^{\alpha n} and H\mathrm{H} being the entropy function, it is Θ(1)min ⁣(1,1(α(1p))(1α)n)2(H(α)+(1α)log2p)n.\Theta(1) \cdot \min\!\left(1, \, \frac{1}{(\alpha\,{-}\,(1-p)) \sqrt{(1\,{-}\,\alpha) n}}\right) \cdot 2^{(\mathrm{H}(\alpha) + (1-\alpha) \log_2 p) n}. This implies that the maximum expected size over all mm is Θ((1+p)n/n)\Theta((1+p)^n/\sqrt{n}). Our structural findings have algorithmic implications for minimizing an input hypergraph, which in turn has applications in the profiling of relational databases as well as for the Orthogonal Vectors problem studied in fine-grained complexity. The main technical tool is an improvement of the Chernoff--Hoeffding inequality, which we make tight up to constant factors. We show that for a binomial variable XBin(n,p)X \sim \mathrm{Bin}(n,p) and real number 0<xp0 < x \le p, it holds that P[Xxn]=Θ(1)min ⁣(1,1(px)xn)2 ⁣D(xp)n\mathrm{P}[X \le xn] = \Theta(1) \cdot \min\!\left(1, \, \frac{1}{(p-x) \sqrt{xn}}\right) \cdot 2^{-\!\mathrm{D}(x \,{\|}\, p) n}, where D\mathrm{D} denotes the Kullback--Leibler divergence between Bernoulli distributions. The result remains true if xx depends on nn as long as it is bounded away from 00.

Keywords

Cite

@article{arxiv.1910.00308,
  title  = {The Minimization of Random Hypergraphs},
  author = {Thomas Bläsius and Tobias Friedrich and Martin Schirneck},
  journal= {arXiv preprint arXiv:1910.00308},
  year   = {2020}
}

Comments

28 pages, 2 figures; Changes: binomial characterization unified, improvement of the Chernoff-Hoeffding theorem extended to case x --> p

R2 v1 2026-06-23T11:31:25.581Z