3

Akbari, Ghodrati, Hosseinzadeh (2017), On the structure of graphs having a unique k-factor, Aust. J. Combin. (pdf) show:

... we prove that there is no r-regular graph (r≥2) with a unique perfect matching.

It seems natural to explore the stronger case:

Question: Does there exist an r-regular graph (r≥2) with a unique maximum matching?

I.e., the same problem with "perfect" replaced by "maximum", thereby including graphs without perfect matchings.

The answer is clearly "no" for r=2 (such graphs are unions of cycles; we can rotate any cycle to give a distinct maximum matching). For r=3, the answer is "no" for bridgeless cubic graphs, which have a perfect matching (by Petersen's Theorem) and therefore the above result applies. I'm not sure beyond this.

Rebecca J. Stones
  • 1,327
  • 7
  • 22

1 Answers1

6

Take a maximum matching $M$ and a vertex $v$ not in $M$. If $v$ has a neighbour $w$ not in $M$, then $M+vw$ is a larger matching. So $v$ has a neighbour $x$ which is in an edge $xy$ of $M$. Now $M-xy+vx$ is another maximum matching.

Brendan McKay
  • 37,203