On Computing the Hamiltonian Index of Graphs

TL;DR

This paper investigates the computational complexity of the Hamiltonian Index of graphs, providing fixed-parameter algorithms based on treewidth and connecting it to the Eulerian Steiner Subgraph problem.

Contribution

It introduces a fixed-parameter algorithm for computing the Hamiltonian Index using treewidth and relates it to the Eulerian Steiner Subgraph problem.

Findings

The Hamiltonian Index problem is NP-hard for fixed k.

A fixed-parameter algorithm for the Hamiltonian Index based on treewidth.

An efficient algorithm for the Eulerian Steiner Subgraph problem.

Abstract

The $r$-th iterated line graph $L^{r}(G)$ of a graph $G$ is defined by: (i) $L^{0}(G) = G$ and (ii) $L^{r}(G) = L(L^{(r- 1)}(G))$ for $r > 0$, where $L(G)$ denotes the line graph of $G$. The Hamiltonian Index $h(G)$ of $G$ is the smallest $r$ such that $L^{r}(G)$ has a Hamiltonian cycle. Checking if $h(G) = k$ is NP-hard for any fixed integer $k \geq 0$ even for subcubic graphs $G$. We study the parameterized complexity of this problem with the parameter treewidth, $tw(G)$, and show that we can find $h(G)$ in time $O*((1 + 2^{(\omega + 3)})^{tw(G)})$ where $\omega$ is the matrix multiplication exponent and the $O*$ notation hides polynomial factors in input size. The NP-hard Eulerian Steiner Subgraph problem takes as input a graph $G$ and a specified subset $K$ of terminal vertices of $G$ and asks if $G$ has an Eulerian (that is: connected, and with all vertices of even degree.) subgraph $H$ containing all the terminals. A second result (and a key ingredient of our algorithm for finding $h(G)$) in this work is an algorithm which solves Eulerian Steiner Subgraph in $O*((1 + 2^{(\omega + 3)})^{tw(G)})$ time.