On sizes of 1-cross intersecting set pair systems

TL;DR

This paper proves an improved upper bound on the maximum size of 1-cross intersecting set pair systems with bounded set sizes, confirming Holzman's conjecture that the bound can be tightened from 29/30 to 5/6.

Contribution

It establishes a tighter upper bound of
6/30 on the size of such systems, advancing understanding of their combinatorial limits.

Findings

Proves that m(a,b,1) b1 bc bc b6 inom{a+b}{a}

Confirms Holzman's conjecture on the upper bound factor

Provides insights into the structure of 1-cross intersecting set pair systems.

Abstract

Let $\{(A_i,B_i)\}_{i=1}^m$ be a set pair system. F\"{u}redi, Gy\'{a}rf\'{a}s and Kir\'{a}ly called it {\em $1$-cross intersecting} if $|A_i\cap B_j|$ is $1$ when $i\neq j$ and $0$ if $i=j$. They studied such systems and their generalizations, and in particular considered $m(a,b,1)$ -- the maximum size of a $1$-cross intersecting set pair system in which $|A_i|\leq a$ and $|B_i|\leq b$ for all $i$. F\"{u}redi, Gy\'{a}rf\'{a}s and Kir\'{a}ly proved that $m(n,n,1)\geq 5^{(n-1)/2}$ and asked whether there are upper bounds on $m(n,n,1)$ significantly better than the classical bound ${2n\choose n}$ of Bollob\' as for cross intersecting set pair systems. Answering one of their questions, Holzman recently proved that if $a,b\geq 2$, then $m(a,b,1)\leq \frac{29}{30}\binom{a+b}{a}$. He also conjectured that the factor $\frac{29}{30}$ in his bound can be replaced by $\frac{5}{6}$. The goal of this paper is to prove this bound.