Complexity of Steiner Tree in Split Graphs - Dichotomy Results

TL;DR

This paper explores the computational complexity of the Steiner tree problem in split graphs, establishing a clear boundary between polynomial-time solvability and NP-completeness based on the graph's forbidden induced subgraphs.

Contribution

It presents a dichotomy result showing Steiner tree is polynomial-time solvable in $K_{1,4}$-free split graphs but NP-complete in $K_{1,5}$-free split graphs, along with structural insights and algorithms.

Findings

Steiner tree is polynomial-time solvable in $K_{1,4}$-free split graphs.

Steiner tree is NP-complete in $K_{1,5}$-free split graphs.

Structural properties of $K_{1,4}$-free and $K_{1,3}$-free split graphs are characterized.

Abstract

Given a connected graph $G$ and a terminal set $R \subseteq V(G)$, {\em Steiner tree} asks for a tree that includes all of $R$ with at most $r$ edges for some integer $r \geq 0$. It is known from [ND12,Garey et. al \cite{steinernpc}] that Steiner tree is NP-complete in general graphs. {\em Split graph} is a graph which can be partitioned into a clique and an independent set. K. White et. al \cite{white} has established that Steiner tree in split graphs is NP-complete. In this paper, we present an interesting dichotomy: we show that Steiner tree on $K_{1,4}$-free split graphs is polynomial-time solvable, whereas, Steiner tree on $K_{1,5}$-free split graphs is NP-complete. We investigate $K_{1,4}$-free and $K_{1,3}$-free (also known as claw-free) split graphs from a structural perspective. Further, using our structural study, we present polynomial-time algorithms for Steiner tree in $K_{1,4}$-free and $K_{1,3}$-free split graphs. Although, polynomial-time solvability of $K_{1,3}$-free split graphs is implied from $K_{1,4}$-free split graphs, we wish to highlight our structural observations on $K_{1,3}$-free split graphs which may be used in other combinatorial problems.