Optimal Algorithm for Paired-Domination in Distance-Hereditary Graphs
Ta-Yu Mu, Ching-Chi Lin
TL;DR
The paper addresses the minimum paired-domination problem on distance-hereditary graphs and presents an optimal $O(n+m)$-time algorithm, with a further $O(n)$-time improvement when a decomposition tree of $G$ is provided. The core approach combines a linear-time decomposition-tree construction with a bottom-up dynamic programming framework that tracks carefully defined subproblem parameters through recursive graph-assembly operations (true twin, false twin, and attachment). Key contributions include two recurrences (Eq. 1 and Eq. 2) and the introduction of auxiliary quantities such as $oldsymbol{ umy{min}}(v)$, $oldsymbol{ umy{ he alpha}}(v)$, $oldsymbol{ umy{ he beta}}(v)$, and the mty indicators, all computed in $O(1)$ per internal node. The results yield a linear-time algorithm for the minimum paired-dominating set on distance-hereditary graphs, with potential extensions to circle- and planar-graph families through related structural decompositions and approximation strategies. This work significantly improves prior $O(n^2)$ bounds and provides a framework for efficient exact solutions in structured graph classes, with practical implications for security and surveillance applications that rely on paired-dominating configurations.
Abstract
The domination problem and its variants represent a classical domain within algorithmic graph theory. Among these variants, the paired-domination problem holds particular prominence due to its real-world implications in security and surveillance domains. Given an input graph $G$, the paired-domination problem involves identifying a minimum dominating set $D$ that induces a subgraph of $G$ with a perfect matching. Lin et al.~[\emph{Paired-domination problem on distance-hereditary graphs}, Algorithmica, 2020] previously presented a solution to this problem with a time complexity of $O(n^2)$. This paper significantly enhances their findings by introducing an $O(n+m)$-time algorithm. Furthermore, the time complexity of this algorithm can be reduced to $O(n)$ when provided with a decomposition tree for the graph $G$.
