English

Kruskal-Katona type Problem

Combinatorics 2018-05-02 v1

Abstract

The Kruskal Katona theorem was proved in the 1960s. In the theorem, we are given an integer rr and families of sets AN(r)\mathcal{A}\subset \mathbb{N}^{(r)} and BN(r1)\mathcal{B}\subset\mathbb{N}^{(r-1)} such that for every AAA\in\mathcal{A}, every subset of AA of size r1r-1 is in B\mathcal{B}. We are interested in finding the mimimum size of b=Bb=|\mathcal{B}| given fixed values of rr and a=Aa=|\mathcal{A}|. The Kruskal Katona theorem states that this mimimum occurs when both A\mathcal{A} and B\mathcal{B} are initial segments of the colexicographic ordering. The Kruskal Katona theorem is very useful and has had many applications and generalisations. In this paper, we are interested in one particular generalisation, where instead of every subset of AA of size r1r-1 being in B\mathcal{B}, we will instead ask that only kk of them are, where kk is some integer smaller than rr. Note that setting k=rk=r is exactly the Kruskal Katona theorem. We will first find exact results for the cases where 0k30\leq k \leq 3. For k4k\geq 4 we will not solve the question completely, however, we will find the exact result for infinitely many aa. We will also provide a formula that is within some additive constant of the correct result for all aa.

Keywords

Cite

@article{arxiv.1805.00340,
  title  = {Kruskal-Katona type Problem},
  author = {Matthew Fitch},
  journal= {arXiv preprint arXiv:1805.00340},
  year   = {2018}
}

Comments

28 pages

R2 v1 2026-06-23T01:41:35.392Z