Table of Contents
Fetching ...

General Strong Bound on the Uncrossed Number which is Tight for the Edge Crossing Number

Gaspard Charvy, Tomáš Masařík

TL;DR

This work establishes a strong, general lower bound on the uncrossed number $unc(G)$ and a matching upper bound on the maximum uncrossed subgraph number $h(G)$, linking these concepts to graph thickness and outerthickness. The main result bounds $h(G)$ by $3n-6 - \sqrt{2m} + \sqrt{6(n-2)}$, which translates, in dense regimes, to tight asymptotics and yields a sharp lower bound on $unc(G)$ via $unc(G) \ge \lceil m / (3n-6-\sqrt{2m}+\sqrt{6(n-2)}) \rceil$. A construction based on wheel graphs demonstrates tightness up to lower-order terms for all densities, while a strengthened bound is derived for triangle-free graphs, together with a triangle-free tightness construction. The paper also posits a tight conjecture for triangle-free graphs, $h(G) \le 2n-4 - \sqrt{m} + \sqrt{2(n-2)}$, inviting future resolution of the density gap and the precise behavior of $unc(G)$ versus $\theta(G)$ and $\theta_o(G)$ in dense regimes.

Abstract

We investigate a very recent concept for visualizing various aspects of a graph in the plane using a collection of drawings introduced by Hliněný and Masařík [GD 2023]. Formally, given a graph $G$, we aim to find an uncrossed collection containing drawings of $G$ in the plane such that each edge of $G$ is not crossed in at least one drawing in the collection. The uncrossed number of $G$ ($unc(G)$) is the smallest integer $k$ such that an uncrossed collection for $G$ of size $k$ exists. The uncrossed number is lower-bounded by the well-known thickness, which is an edge-decomposition of $G$ into planar graphs. This connection gives a trivial lower-bound $\lceil\frac{|E(G)|}{3|V(G)|-6}\rceil \le unc(G)$. In a recent paper, Balko, Hliněný, Masařík, Orthaber, Vogtenhuber, and Wagner [GD 2024] presented the first non-trivial and general lower-bound on the uncrossed number. We summarize it in terms of dense graphs (where $|E(G)|=ε(|V(G)|)^2$ for some $ε>0$): $\lceil\frac{|E(G)|}{c_ε|V(G)|}\rceil \le unc(G)$, where $c_ε\ge 2.82$ is a constant depending on $ε$. We improve the lower-bound to state that $\lceil\frac{|E(G)|}{3|V(G)|-6-\sqrt{2|E(G)|}+\sqrt{6(|V(G)|-2)}}\rceil \le unc(G)$. Translated to dense graphs regime, the bound yields a multiplicative constant $c'_ε=3-\sqrt{(2-ε)}$ in the expression $\lceil\frac{|E(G)|}{c'_ε|V(G)|+o(|V(G)|)}\rceil \le unc(G)$. Hence, it is tight (up to low-order terms) for $ε\approx \frac{1}{2}$ as warranted by complete graphs. In fact, we formulate our result in the language of the maximum uncrossed subgraph number, that is, the maximum number of edges of $G$ that are not crossed in a drawing of $G$ in the plane. In that case, we also provide a construction certifying that our bound is asymptotically tight (up to low-order terms) on dense graphs for all $ε>0$.

General Strong Bound on the Uncrossed Number which is Tight for the Edge Crossing Number

TL;DR

This work establishes a strong, general lower bound on the uncrossed number and a matching upper bound on the maximum uncrossed subgraph number , linking these concepts to graph thickness and outerthickness. The main result bounds by , which translates, in dense regimes, to tight asymptotics and yields a sharp lower bound on via . A construction based on wheel graphs demonstrates tightness up to lower-order terms for all densities, while a strengthened bound is derived for triangle-free graphs, together with a triangle-free tightness construction. The paper also posits a tight conjecture for triangle-free graphs, , inviting future resolution of the density gap and the precise behavior of versus and in dense regimes.

Abstract

We investigate a very recent concept for visualizing various aspects of a graph in the plane using a collection of drawings introduced by Hliněný and Masařík [GD 2023]. Formally, given a graph , we aim to find an uncrossed collection containing drawings of in the plane such that each edge of is not crossed in at least one drawing in the collection. The uncrossed number of () is the smallest integer such that an uncrossed collection for of size exists. The uncrossed number is lower-bounded by the well-known thickness, which is an edge-decomposition of into planar graphs. This connection gives a trivial lower-bound . In a recent paper, Balko, Hliněný, Masařík, Orthaber, Vogtenhuber, and Wagner [GD 2024] presented the first non-trivial and general lower-bound on the uncrossed number. We summarize it in terms of dense graphs (where for some ): , where is a constant depending on . We improve the lower-bound to state that . Translated to dense graphs regime, the bound yields a multiplicative constant in the expression . Hence, it is tight (up to low-order terms) for as warranted by complete graphs. In fact, we formulate our result in the language of the maximum uncrossed subgraph number, that is, the maximum number of edges of that are not crossed in a drawing of in the plane. In that case, we also provide a construction certifying that our bound is asymptotically tight (up to low-order terms) on dense graphs for all .

Paper Structure

This paper contains 9 sections, 20 theorems, 66 equations, 6 figures.

Key Result

Theorem 1.1

Every connected graph $G$ with $n \geq 3$ vertices and $m \geq 0$ edges satisfies

Figures (6)

  • Figure 1: Multiplicative constant $c_\varepsilon$ given by \ref{['thm:old-lb']} (Uncrossed2) depected by dashed blue line and \ref{['eq:dense']} drawn as full red line. Note that our values of $c_\varepsilon$ are tight for $\varepsilon$ tending towards $\frac{1}{2}$ (complete graphs) and towards 0 (planar triangulated graphs).
  • Figure 2: The wheel graph on six vertices $W_6$; see \ref{['sec:prelim']} for the formal definition.
  • Figure 3: An example of $DG_{5,11}$. Crossed edges are depicted using dotted lines. See the \ref{['constr']} paragraph for the formal description.
  • Figure 4: A visual representation of the asymptotical intersection between $\frac{r_3}{n}$ (red line) and $\frac{b_3^{\pm}}{n}$ for $\varepsilon = 0.5$ (blue line). The filled area (in gray) is highlighting the possible values of $\frac{m_i}{n}$. The maximum possible value of $\frac{m_i}{n}$ corresponds to the multiplicative constant $c_\varepsilon$ described in the introduction.
  • Figure 5: An example of $DK_{4,4}$. See the \ref{['constr_DKxx']} paragraph for the formal description.
  • ...and 1 more figures

Theorems & Definitions (33)

  • Theorem 1.1: Uncrossed2
  • Theorem 1.2: Exact $h$ value for complete ringel64 and complete bipartite germanPaper graphs
  • Theorem 1.3: Uncrossed Number Lower Bound
  • Theorem 1.4: Uncrossed Number Lower Bound for Triangle-free Graphs
  • Theorem 1.6: Maximum Uncrossed Subgraph Number Upper Bound
  • Theorem 1.7: Maximum Uncrossed Subgraph Tight Construction
  • Theorem 1.8: Maximum Uncrossed Subgraph Number Upper Bound for Triangle-Free Graphs
  • Theorem 1.9: Maximum Uncrossed Subgraph Construction for Triangle-Free Graphs
  • Theorem 2.1: Maximum Uncrossed Subgraph Tight Construction
  • Theorem 2.1: ringel64
  • ...and 23 more