[Paper Review] Linear embeddings of $K_9$ are triple linked

Ramin Naimi, Elena Pavelescu|arXiv (Cornell University)|Feb 8, 2012

Advanced Graph Theory Research6 references3 citations

TL;DR

This paper proves that every linear embedding of the complete graph $K_9$ in 3D space contains a non-split 3-component link, using oriented matroid theory to systematically analyze all acyclic oriented matroids of rank 4 on 9 elements. The authors compute over 2.3 billion acyclic oriented matroids, confirming that each contains a triple link, thereby disproving a potential extension of Sachs’ conjecture to 3-links.

ABSTRACT

We use the theory of oriented matroids to show that any linear embedding of $K_9$, the complete graph on nine vertices, contains a non-split link with three components.

Motivation & Objective

• To determine whether every linear embedding of $K_9$ in $\mathbb{R}^3$ contains a non-split 3-component link.
• To test the validity of an extension of Sachs’ conjecture on linkless embeddings to 3-links.
• To use oriented matroid theory to systematically analyze the linking structure of linear $K_9$ embeddings.
• To verify computationally that no linear $K_9$ embedding avoids non-split 3-links, despite $K_9$ not being intrinsically 3-linked in general.

Proposed method

• Represent each linear embedding of $K_9$ via an oriented matroid of rank 4 on 9 elements, derived from vertex coordinates and chirotope sign patterns.
• Use the database of all isomorphism classes of uniform oriented matroids $\mathrm{OM}(9,4)$, which includes over 2.3 billion oriented matroids.
• For each acyclic oriented matroid, compute the set of reorientation sets $A$ for which $\mathcal{M}_{-A}$ is cyclic, using Lemma 4.3.
• For each acyclic oriented matroid, identify all 3–2–partitioned circuits to determine which edges pierce triangles, enabling link detection via Lemma 4.1.
• Apply a computational algorithm in Mathematica to check all possible disjoint triangle triples, using the linking condition defined in Definition 4.2.
• Verify results via supporting evidence: recompute known results for $K_6$ and $K_7$, and manually verify three random isomorphism classes of $\mathrm{OM}(9,4)$.

Experimental results

Research questions

• RQ1 Does every linear embedding of $K_9$ in $\mathbb{R}^3$ contain a non-split 3-component link?
• RQ2 Can Sachs’ conjecture on linear linkless embeddings be extended to 3-links?
• RQ3 Is there a linear embedding of $K_9$ that avoids all non-split 3-links, despite $K_9$ not being intrinsically 3-linked in general?
• RQ4 Can oriented matroid theory and computational enumeration reliably detect non-split 3-links in spatial graphs?
• RQ5 Are the results of the computation verifiable through independent manual checks and documented evidence?

Key findings

• All 2,374,808,320 acyclic oriented matroids in $\mathrm{OM}(9,4)$ contain at least one non-split 3-component link.
• The computation confirmed that every acyclic oriented matroid in $\mathrm{OM}(9,4)$ contains a triple link, implying all linear $K_9$ embeddings are triple linked.
• The program’s results for $K_6$ and $K_7$ matched known results from prior literature, validating the computational approach.
• Supporting evidence includes detailed output for three randomly selected isomorphism classes of $\mathrm{OM}(9,4)$, with full manual verification possible for any listed 3-link.
• The authors provide a fully documented Mathematica program on arXiv, enabling independent verification of all computations.
• Despite the strong computational evidence, the authors note that not all acyclic oriented matroids in $\mathrm{OM}(9,4)$ are necessarily induced by actual linear embeddings of $K_9$.