Tur\'an number of special four cycles in triple systems

TL;DR

This paper determines the asymptotic maximum size of triple systems avoiding certain special four-cycle configurations, extending known bounds for Turán numbers related to Berge cycles.

Contribution

It establishes the order of magnitude for Turán numbers of special four-cycle configurations in triple systems, including several new cases and extending previous results.

Findings

Turán number for special four-cycles is Θ(n^{3/2})

Identified cases where Turán number is Θ(n^{3/2}) or Θ(n^{2})

Most cases remain unsolved, indicating ongoing challenges

Abstract

A {\em special four-cycle } $F$ in a triple system consists of four triples {\em inducing } a $C_4$. This means that $F$ has four special vertices $v_1,v_2,v_3,v_4$ and four triples in the form $w_iv_iv_{i+1}$ (indices are understood $\pmod 4$) where the $w_j$s are not necessarily distinct but disjoint from $\{v_1,v_2,v_3,v_4\}$. There are seven non-isomorphic special four-cycles, their family is denoted by $\cal{F}$. Our main result implies that the Tur\'an number $\text{ex}(n,{\cal{F}})=\Theta(n^{3/2})$. In fact, we prove more, $\text{ex}(n,\{F_1,F_2,F_3\})=\Theta(n^{3/2})$, where the $F_i$-s are specific members of $\cal{F}$. This extends previous bounds for the Tur\'an number of triple systems containing no Berge four cycles. We also study $\text{ex}(n,{\cal{A}})$ for all ${\cal{A}}\subseteq {\cal{F}}$. For 16 choices of $\cal{A}$ we show that $\text{ex}(n,{\cal{A}})=\Theta(n^{3/2})$, for 92 choices of $\cal{A}$ we find that $\text{ex}(n,{\cal{A}})=\Theta(n^2)$ and the other 18 cases remain unsolved.