On Kotzig's conjecture in random graphs

Abstract

In 1963, Anton Kotzig famously conjectured that $K_{n}$, the complete graph of order $n$, where $n$ is even, can be decomposed into $n-1$ perfect matchings such that every pair of these matchings forms a Hamilton cycle. The problem is still wide open and here we consider a variant of it for the binomial random graph $G(n,p)$. We prove that, for every fixed $k$, there exists a constant $C=C(k)$ such that, when $p\ge \frac{C \log n}{n}$, with high probability, $G(n,p)$ contains $k$ edge-disjoint perfect matchings with the property that every pair of them forms a Hamilton cycle. In fact, our main result is a very precise counting result for $K_n$. We show that, given any $k$ edge-disjoint perfect matchings $M_1,\dots,M_k$, the probability that a uniformly random perfect matching $M^*$ in $K_n$ has the property that $M^*\cup M_i$ forms a Hamilton cycle for each $i\in [k]$ is $Θ_k(n^{-k/2})$. This is proved by building on a variety of methods, including a random process analysis, the absorption method, the entropy method and the switching method. The result on the binomial random graph follows from a slight strengthening of our counting result via the recent breakthroughs on the expectation threshold conjecture.