Parameterized algorithms for $k$-Inversion

TL;DR

This paper develops fixed-parameter tractable algorithms for the $k$-Inversion problem on various graph classes, including tournaments, block graphs, and graphs with bounded treewidth, extending previous work.

Contribution

It generalizes existing FPT algorithms for tournaments to broader graph classes and introduces a new FPT algorithm based on treewidth for general directed graphs.

Findings

FPT algorithm for $k$-Inversion on tournaments and their variants.

FPT algorithm for $k$-Inversion on block graphs.

Algorithm with runtime $2^{O( ext{tw}(k+ ext{tw}))} imes n^{O(1)}$ for general graphs.

Abstract

Inversion of a directed graph $D$ with respect to a vertex subset $Y$ is the directed graph obtained from $D$ by reversing the direction of every arc whose endpoints both lie in $Y$. More generally, the inversion of $D$ with respect to a tuple $(Y_1, Y_2, \ldots, Y_\ell)$ of vertex subsets is defined as the directed graph obtained by successively applying inversions with respect to $Y_1, Y_2, \ldots, Y_\ell$. Such a tuple is called a \emph{decycling family} of $D$ if the resulting graph is acyclic. In the \textsc{$k$-Inversion} problem, the input consists of a directed graph $D$ and an integer $k$, and the task is to decide whether $D$ admits a decycling family of size at most $k$. Alon et al.\ (SIAM J.\ Discrete Math., 2024) proved that the problem is NP-complete for every fixed value of $k$, thereby ruling out XP algorithms, and presented a fixed-parameter tractable (FPT) algorithm parameterized by $k$ for tournament inputs. In this paper, we generalize their algorithm to a broader variant of the problem on tournaments and subsequently use this result to obtain an FPT algorithm for \textsc{$k$-Inversion} when the underlying undirected graph of the input is a block graph. Furthermore, we obtain an algorithm for \textsc{$k$-Inversion} on general directed graphs with running time $2^{O(\mathrm{tw}(k + \mathrm{tw}))} \cdot n^{O(1)}$, where $\mathrm{tw}$ denotes the treewidth of the underlying graph.