The Complexity of Maximum $k$-Order Bounded Component Set Problem (1712.02870v3)

Abstract: Given a graph $G=(V, E)$ and a positive integer $k$, in Maximum $k$-Order Bounded Component Set (Max-$k$-OBCS), it is required to find a vertex set $S \subseteq V$ of maximum size such that each component in the induced graph $G[S]$ has at most $k$ vertices. We prove that for constant $k$, Max-$k$-OBCS is hard to approximate within a factor of $n^{1 -\epsilon}$, for any $\epsilon > 0$, unless $\mathsf{P} = \mathsf{NP}$. This is an improvement on the previous lower bound of $\sqrt{n}$ for Max-2-OBCS due to Orlovich et al. We provide lower bounds on the approximability when $k$ is not a constant as well. Max-$k$-OBCS can be seen as a generalization of Maximum Independent Set (Max-IS). We generalize Tur\'an's greedy algorithm for Max-IS and prove that it approximates Max-$k$-OBCS within a factor of $(2k - 1)\overline{d} + k$, where $\overline{d}$ is the average degree of the input graph $G$. This approximation factor is a generalization of Tur\'an's approximation factor for Max-IS.