On the structure of a smallest counterexample and a new class verifying the 2-Decomposition Conjecture

TL;DR

This paper investigates the 2-Decomposition Conjecture in graph theory, analyzing the structure of minimal counterexamples and proving the conjecture for specific graph classes, advancing understanding towards a general proof.

Contribution

It characterizes structural properties of minimal counterexamples and proves the conjecture for graphs with certain cactus-like structures of degree-3 vertices.

Findings

Minimum counterexamples have girth at least 5.

Vertices of degree 2 in minimal counterexamples are at least 3 edges apart.

The conjecture holds for graphs where degree-3 vertices form a collection of cycles.

Abstract

The 2-Decomposition Conjecture, equivalent to the 3-Decomposition Conjecture stated in 2011 by Hoffmann-Ostenhof, claims that every connected graph $G$ with vertices of degree 2 and 3, for which $G \setminus E(C)$ is disconnected for every cycle $C$, admits a decomposition into a spanning tree and a matching. In this work we present two main results focused on developing a strategy to prove the 2-Decomposition Conjecture. One of them is a list of structural properties of a minimum counterexample for this conjecture. Among those properties, we prove that a minimum counterexample has girth at least 5 and its vertices of degree 2 are at distance at least 3. Motivated by the class of smallest counterexamples, we show that the 2-Decomposition Conjecture holds for graphs whose vertices of degree 3 induce a collection of cacti in which each vertex belongs to a cycle. The core of the proof of this result may possibly be used in an inductive proof of the 2-Decomposition Conjecture based on a parameter that relates the number of vertices of degree 2 and 3 in a minimum counterexample.