On the Parameterized Complexity of Eulerian Strong Component Arc Deletion

TL;DR

This paper investigates the computational complexity of the Eulerian Strong Component Arc Deletion problem, establishing its hardness and identifying conditions under which it admits fixed-parameter tractable algorithms.

Contribution

It provides the first comprehensive complexity analysis of the problem, including hardness results and algorithms for various parameterizations.

Findings

The problem is W[1]-hard when parameterized by solution size.

It is in XP when parameterized by treewidth.

It is fixed-parameter tractable when parameterized by combined parameters like treewidth and maximum degree.

Abstract

In this paper, we study the Eulerian Strong Component Arc Deletion problem, where the input is a directed multigraph and the goal is to delete the minimum number of arcs to ensure every strongly connected component of the resulting digraph is Eulerian. This problem is a natural extension of the Directed Feedback Arc Set problem and is also known to be motivated by certain scenarios arising in the study of housing markets. The complexity of the problem, when parameterized by solution size (i.e., size of the deletion set), has remained unresolved and has been highlighted in several papers. In this work, we answer this question by ruling out (subject to the usual complexity assumptions) a fixed-parameter tractable (FPT) algorithm for this parameter and conduct a broad analysis of the problem with respect to other natural parameterizations. We prove both positive and negative results. Among these, we demonstrate that the problem is also hard (W[1]-hard or even para-NP-hard) when parameterized by either treewidth or maximum degree alone. Complementing our lower bounds, we establish that the problem is in XP when parameterized by treewidth and FPT when parameterized either by both treewidth and maximum degree or by both treewidth and solution size. We show that these algorithms have near-optimal asymptotic dependence on the treewidth assuming the Exponential Time Hypothesis.