📝 SummarXiv

Abstract

For a graph $G$ and a parameter $k$, we call a vertex $k$-enabling if it belongs both to a clique of size $k$ and to an independent set of size $k$, and we call it $k$-excluding otherwise. Motivated by issues that arise in secret sharing schemes, we study the complexity of detecting vertices that are $k$-excluding. We show that for every $\epsilon$, for sufficiently large $n$, if $k > (\frac{1}{4} + \epsilon)n$, then every graph on $n$ vertices must have a $k$-excluding vertex, and moreover, such a vertex can be found in polynomial time. In contrast, if $k < (\frac{1}{4} - \epsilon)n$, a regime in which it might be that all vertices are $k$-enabling, deciding whether a graph has no $k$-excluding vertex is NP-hard.