Constructing Hamiltonian Decompositions of Complete $k$-Uniform Hypergraphs
Javad Maheri, Petros Elia
TL;DR
The paper addresses the problem of partitioning the edge set of the complete $k$-uniform hypergraph $K_n^k$ into Hamiltonian cycles, extending previous existence results to explicit designs for all $k$ with prime $n$ and $k< n/2$. It introduces a generator-based framework that maps each $k$-subset to a Hamiltonian cycle via a generator vector $oldsymbol{g}$, a representative vector $oldsymbol{r}_{oldsymbol{g}}$, and the cycle $oldsymbol{ ext{S}}_{oldsymbol{g}}$, with a cycle period equal to the number of vertices $n$ when $ abla$ conditions hold (e.g., $ ext{gcd}(n,k)=1$). By constructing a comprehensive set of generators $oldsymbol{G}_{ ext{tot}}$, determined from parameters $oldsymbol{ extsigma}_{ ext{min}}=k-1$ and $oldsymbol{ extsigma}_{ ext{max}}=n-ig\lceil n/kig ceil$, the authors ensure that the corresponding vector sets partition $oldsymbol{A}_{n,k}$ into $N=inom{n}{k}/n$ disjoint Hamiltonian cycles, providing an explicit Hamiltonian decomposition. The work delivers concrete constructions for all $k$ with prime $n$, with a detailed example for $(n,k)=(11,5)$ and discussion of practical implications for distributed systems, coded caching, routing, and fault tolerance.
Abstract
Motivated by the wide-ranging applications of Hamiltonian decompositions in distributed computing, coded caching, routing, resource allocation, load balancing, and fault tolerance, our work presents a comprehensive design for Hamiltonian decompositions of complete $k$-uniform hypergraphs $K_n^k$. Building upon the resolution of the long-standing conjecture of the existence of Hamiltonian decompositions of complete hypergraphs, a problem that was resolved using existence-based methods, our contribution goes beyond the previous explicit designs, which were confined to the specific cases of $k=2$ and $k=3$, by providing explicit designs for all $k$ and $n$ prime, allowing for a broad applicability of Hamiltonian decompositions in various settings.
