Table of Contents
Fetching ...

Dirac-type condition for Hamilton-generated graphs

Xinmin Hou, Zhi Yin

TL;DR

This work resolves a Dirac-type problem for Hamilton-generated graphs by showing that for sufficiently large odd $n$, every Hamilton-connected graph on $n$ vertices with minimum degree at least $\frac{n-1}{2}$ satisfies $\mathcal{C}_{n}(G)=\mathcal{C}(G)$, i.e., every cycle is a symmetric difference of Hamilton cycles. The authors deploy the parity-switcher technique (from recent work of Christoph, Nenadov, and Petrova) together with a robust structural classification to rule out all potential obstructions to Hamilton-generatedness. The result is tight in both the minimum-degree bound and the Hamilton-connected condition, with constructive examples illustrating the limits. This advances Bondy-type questions under Dirac-type thresholds and suggests the parity-switcher framework as a powerful tool for cycle-space generation problems in dense graphs.

Abstract

The cycle space $\mathcal{C}(G)$ of a graph $G$ is defined as the linear space spanned by all cycles in $G$. For an integer $k\ge 3$, let $\mathcal{C}_k (G)$ denote the subspace of $\mathcal{C}(G)$ generated by the cycles of length exactly $k$. A graph $G$ on $n$ vertices is called Hamilton-generated if $\mathcal{C}_n (G) = \mathcal{C}(G)$, meaning every cycle in $G$ is a symmetric difference of some Hamilton cycles of $G$. %A necessary condition for this property is that $n$ must be odd. Heinig (European J. Combin., 2014) showed that for any $σ>0$ and sufficiently large odd $n$, every $n$-vertex graph with minimum degree $(1+ σ)n/2$ is Hamilton-generated. He further posed the question that whether the minimum degree requirement could be lowered to the Dirac threshold $n/2$. Recent progress by Christoph, Nenadov, and Petrova~(arXiv:2402.01447) reduced the minimum degree condition to $n/2 + C$ for some large constant $C$. In this paper, we resolve Heinig's problem completely by proving that for sufficiently large odd $n$, every Hamilton-connected graph $G$ on $n$ vertices with minimum degree at least $(n-1)/2$ is Hamilton-generated. Moreover, this result is tight for the minimum degree and the Hamilton-connected condition. The proof relies on the parity-switcher technique introduced by Christoph, et al in their recent work, as well as a classification lemma that strengthens a previous result by Krivelevich, Lee, and Sudakov~(Trans. Amer. Math. Soc., 2014).

Dirac-type condition for Hamilton-generated graphs

TL;DR

This work resolves a Dirac-type problem for Hamilton-generated graphs by showing that for sufficiently large odd , every Hamilton-connected graph on vertices with minimum degree at least satisfies , i.e., every cycle is a symmetric difference of Hamilton cycles. The authors deploy the parity-switcher technique (from recent work of Christoph, Nenadov, and Petrova) together with a robust structural classification to rule out all potential obstructions to Hamilton-generatedness. The result is tight in both the minimum-degree bound and the Hamilton-connected condition, with constructive examples illustrating the limits. This advances Bondy-type questions under Dirac-type thresholds and suggests the parity-switcher framework as a powerful tool for cycle-space generation problems in dense graphs.

Abstract

The cycle space of a graph is defined as the linear space spanned by all cycles in . For an integer , let denote the subspace of generated by the cycles of length exactly . A graph on vertices is called Hamilton-generated if , meaning every cycle in is a symmetric difference of some Hamilton cycles of . %A necessary condition for this property is that must be odd. Heinig (European J. Combin., 2014) showed that for any and sufficiently large odd , every -vertex graph with minimum degree is Hamilton-generated. He further posed the question that whether the minimum degree requirement could be lowered to the Dirac threshold . Recent progress by Christoph, Nenadov, and Petrova~(arXiv:2402.01447) reduced the minimum degree condition to for some large constant . In this paper, we resolve Heinig's problem completely by proving that for sufficiently large odd , every Hamilton-connected graph on vertices with minimum degree at least is Hamilton-generated. Moreover, this result is tight for the minimum degree and the Hamilton-connected condition. The proof relies on the parity-switcher technique introduced by Christoph, et al in their recent work, as well as a classification lemma that strengthens a previous result by Krivelevich, Lee, and Sudakov~(Trans. Amer. Math. Soc., 2014).

Paper Structure

This paper contains 4 sections, 17 theorems, 171 equations, 1 figure.

Key Result

Theorem 1.2

For any $\sigma >0$ and sufficiently large odd $n$, if $G$ is a graph on $n$ vertices with minimum degree $(1+ \sigma)n/2$, then $\mathcal{C}_n (G) = \mathcal{C}(G)$.

Figures (1)

  • Figure 1: $G_1,G_2,G_3$ in Construction A

Theorems & Definitions (64)

  • Conjecture 1.1: Bondy, 1979
  • Theorem 1.2: Heinig, Heinig
  • Theorem 1.4: Nenadov
  • Theorem 1.5
  • Corollary 1.6
  • Proposition 1
  • proof
  • Lemma 2.1: Nenadov
  • Definition 2.2: Parity-switcher
  • Corollary 2.3
  • ...and 54 more