π€ AI Summary
This paper addresses the piercing problem for families of non-piercing regions in the plane satisfying the $(p,2)$-propertyβi.e., among any $p$ regions, some two intersect. The goal is to determine whether a finite point set can pierce the entire family. Prior work achieved an $O(p^9)$ upper bound on the minimum number of required points. Here, under the broader non-piercing region model, we exploit structural properties of the intersection hypergraph and map the problem to a hypergraph defined on a hereditary linear Delaunay graph. Combining this geometric reduction with hypergraph piercing theory, we provide a concise combinatorial proof that the piercing number is bounded by $O(p)$. This constitutes the first linear upper bound for this setting. Our result not only substantially improves the best-known bound but also unifies and simplifies the proof framework, offering novel insights and key technical tools for the general $(p,q)$-piercing problem.
π Abstract
A family of sets satisfies the $(p,2)$-property if among any $p$ sets in the family, some two intersect. Two recent works used elaborate geometric techniques to show that any family of non-piercing regions in the plane that satisfies the $(p,2)$-property can be pierced by $O(p^9)$ points. In this note we show that even in a much more general setting, piercing by $O(p)$ points can be deduced from known results on hypergraphs with a hereditarily linear Delaunay graph, which include intersection hypergraphs of non-piercing regions.