Lovász matching number in Cayley line hypergraphs
Question 4.4 · arXiv:2505.05339
Status disproved high confidence
Question 4.4 asks whether the line hypergraph of every $r$-regular $r$-edge-colourable Cayley graph must contain $r-1$ vertices whose deletion reduces its matching number. This was answered negatively by Abiad, Garbe, Povill, and Spiegel (arXiv:2506.21286, June 2025), who found a cubic Cayley graph on 36 vertices and a quartic Cayley graph on 96 vertices whose line hypergraphs are counterexamples to Lovász's conjecture, thus disproving the conjecture even in the Cayley graph setting.
Cited literature (1)
-
Provides a cubic Cayley graph on 36 vertices and a quartic Cayley graph on 96 vertices whose line hypergraphs are counterexamples to Lovász's conjecture, directly answering Question 4.4 in the negative.
Reviewer notes. The follow-up paper arXiv:2506.21286 explicitly addresses Question 4.4 from the source paper, noting 'while the generalized Petersen graphs we use for Theorem 2 are not Cayley, we did find a specific example of a cubic Cayley graph on 36 vertices … as well as a quartic Cayley graph on 96 vertices.' This directly disproves the conjecture.
Context
Both known counterexamples to Lovász's conjecture (the Biggs-Smith graph and F168D) are non-Cayley graphs. The authors observe that while symmetry aids in identifying counterexamples, it is unclear why not being a Cayley graph is relevant, motivating this restricted version of Lovász's conjecture.
Source paper
A Counterexample to a Conjecture of Lovász
Alexander Clow, Penny Haxell, Bojan Mohar · 2025-06-10
https://arxiv.org/abs/2505.05339