Bounding connected tree-width

TL;DR

This paper refines bounds on the connected tree-width of graphs, relating it more accurately to tree-width and geodesic cycles, and disproves a conjecture about duality with brambles.

Contribution

It improves the known bounds on connected tree-width and provides a counterexample to a conjectured duality with brambles.

Findings

Improved bound on connected tree-width relative to tree-width and geodesic cycles.

Constructed a graph where connected tree-width exceeds the connected order of any bramble.

Disproved the conjecture of a duality between connected tree-width and brambles.

Abstract

Diestel and M\"uller showed that the connected tree-width of a graph $G$, i.e., the minimum width of any tree-decomposition with connected parts, can be bounded in terms of the tree-width of $G$ and the largest length of a geodesic cycle in $G$. We improve their bound to one that is of correct order of magnitude. Finally, we construct a graph whose connected tree-width exceeds the connected order of any of its brambles. This disproves a conjecture by Diestel and M\"uller asserting an analogue of tree-width duality.