Embedding the Erdős-Rényi Hypergraph into the Random Regular Hypergraph and Hamiltonicity

Abstract: We establish an inclusion relation between two uniform models of random $k$-graphs (for constant $k \ge 2$) on $n$ labeled vertices: $\mathbb G^{(k)}(n,m)$, the random $k$-graph with $m$ edges, and $\mathbb R^{(k)}(n,d)$, the random $d$-regular $k$-graph. We show that if $n\log n\ll m\ll n^k$ we can choose $d = d(n) \sim {km}/n$ and couple $\mathbb G^{(k)}(n,m)$ and $\mathbb R^{(k)}(n,d)$ so that the latter contains the former with probability tending to one as $n\to\infty$. This extends an earlier result of Kim and Vu about "sandwiching random graphs". In view of known threshold theorems on the existence of different types of Hamilton cycles in $\mathbb G^{(k)}(n,m)$, our result allows us to find conditions under which $\mathbb R^{(k)}(n,d)$ is Hamiltonian. In particular, for $k\ge 3$ we conclude that if $n^{k-2} \ll d \ll n^{k-1}$, then a.a.s. $\mathbb R^{(k)}(n,d)$ contains a tight Hamilton cycle.