New results on simplex-clusters in set systems (Q2236657)

Let \(\mathcal{F} = \{A_1, \ldots, A_{d+1}\}\) be a family of \(k\)-subsets of an \(n\)-element set such that the intersection of all members is empty. The family is called a \(d\)-simplex if the intersection of any \(d\) members is non-empty. The family is called a \(d\)-cluster if the union of all members is of size \(\leq 2k\). The family is called a \(d\)-simplex-cluster if it is both a \(d\)-simplex and a \(d\)-cluster. The main result of this paper is a solution of the \textit{P. Keevash} and \textit{D. Mubayi}'s conjecture [Combinatorica 30, No. 2, 175--200 (2010; Zbl 1224.05249)] in the range \(4 \leq d+1 \leq k\) and \(n \geq 2k-d+2\). Suppose that the family \(\mathcal{F}\) contains no \(d\)-simplex-cluster. Then it is proved that the size of \(\mathcal{F}\) is at most \(\binom{n-1}{k-1}\). A family attains the maximum size only when it is a star, i.e., there is an element belonging to all members. A consequence of the main result is that the Erdős-Chvátal \(d\)-simplex conjecture holds for all values of \(d\), \(k\) and \(n\) except for very small values of \(n\).