arXiv reader

A New Upper Bound for the Turán Density of the Tetrahedron

Gyeongwon Jeong, Seonghun Park, Seonghyuk Im, Joonkyung Lee and Hongseok Yang

arXiv
2609.27495v2 · PDF · 2 October 2026
Categories
math.CO, cs.DM, math.OC
Licence
CC BY 4.0 The site shows the text.

Abstract

We prove that the Turán density of the tetrahedron K_4^{(3)} satisfies \pi(K_4^{(3)}) \le 14993367693127837/26880000000000000 < 0.557789, improving Baber’s upper bound of 0.5615 and closing about 62\% of the gap to the conjectured value 5/9. The proof uses an exact seven-vertex flag-algebra certificate incorporating degree-stationarity from Razborov’s differential method. To find the certificate, we combine the established techniques of cutting planes and column generation to optimize jointly over flag families whose types have at most five vertices. We give a complete formal proof of this Turán density bound in Lean 4.

Items 16

Theorem

Theorem 1.1 (main bound)

LY5I

The Turán density of the tetrahedron satisfies

\pi(K_4^{(3)})\;\le\;\frac{14993367693127837}{26880000000000000}\;<\;0.557789.
Lemma

Lemma 2.1 (chain rule )

QUOU

For all F\in\mathcal F^\sigma_m, G\in\mathcal F^\sigma_n and m\le\ell\le n,

p(F;G)\;=\;\sum_{F'\in\mathcal F^\sigma_\ell}p(F;F')\,p(F';G).
Theorem

Theorem 2.2

BQ2N

The elements of \operatorname{Hom}^+(\mathcal A^\sigma,\mathbb R) are exactly the limits of convergent sequences \{G_n\} of \sigma-flags, with \phi(F)=\lim_{n\to\infty}p(F;G_n) for every F\in\mathcal F^\sigma.

Used by Proof
Lemma

Lemma 2.3 (coefficient bound)

FBIW

If f=\sum_{F\in\mathcal F^\sigma_\ell}c_F\,F in \mathcal A^\sigma, then f\le\max_{F\in\mathcal F^\sigma_\ell}c_F.

Used by Proof
Proof of Lemma 2.3

For every \phi\in\operatorname{Hom}^+(\mathcal A^\sigma,\mathbb R), positivity and (?) give

\phi(f)=\sum_{F\in\mathcal F^\sigma_\ell}c_F\phi(F) \;\le\;\Bigl(\max_{F\in\mathcal F^\sigma_\ell}c_F\Bigr) \sum_{F\in\mathcal F^\sigma_\ell}\phi(F) \;=\;\max_{F\in\mathcal F^\sigma_\ell}c_F .\qedhere
Proposition

Proposition 2.4

SJ5G

\pi(K_4^{(3)})=\max_{\phi\in\operatorname{Hom}^+(\mathcal A^0,\mathbb R)}\phi(\rho).

Used by Proof
Proof of Proposition 2.4

For each n\ge3, choose a K_4^{(3)}-free 3-graph G_n on n vertices with \operatorname{ex}(n,K_4^{(3)}) edges. Since \mathcal F^0 is countable and every density lies in [0,1], a diagonal argument gives a subsequence along which p(F;G_n) converges for every F\in\mathcal F^0. By Theorem 2.2, these limits define a positive homomorphism \psi with \psi(\rho)=\pi(K_4^{(3)}). Conversely, every \phi\in\operatorname{Hom}^+(\mathcal A^0,\mathbb R) is realized by a convergent sequence \{G'_n\} with |V(G'_n)|\to\infty. Taking limits in

p(\rho;G'_n)\le \frac{\operatorname{ex}(|V(G'_n)|,K_4^{(3)})}{\binom{|V(G'_n)|}{3}}

gives \phi(\rho)\le\pi(K_4^{(3)}).

Theorem

Theorem 2.5

BJF2

If f\ge0 in \mathcal A^\sigma, then \left\llbracket f\right\rrbracket_\sigma\ge0 in \mathcal A^0.

Used by Proof
Proposition

Proposition 2.6

0QBU

Every maximizer of \phi\mapsto\phi(\rho) on \operatorname{Hom}^+(\mathcal A^0,\mathbb R) is degree-stationary.

Used by Proof
Proof of Proposition 2.6

Let \pi^1\colon\mathcal A^0\to\mathcal A^1 and \partial_1\colon\mathcal A^0\to\mathcal A^1 be Razborov’s upward operator and vertex differential, respectively [Razborov2007, §§2.3.1 and 4.3]. For M\in\mathcal F^0_\ell, \pi^1(M) is the sum of the 1-flags on \ell+1 vertices whose non-root vertices induce a copy of M. For the hyperedge, \partial_1\rho=3(\pi^1(\rho)-d), so [Razborov2007, Theorem 2.8(a)] gives

\left\llbracket d\cdot\partial_1\rho\right\rrbracket_1 =3\bigl(\rho\cdot\left\llbracket d\right\rrbracket_1-\left\llbracket d\cdot d\right\rrbracket_1\bigr)=S .

Razborov calls a theory vertex uniform if it has exactly one one-vertex type [Razborov2007, Definition 11]. This condition holds for \mathcal{T}, since the only 3-graph on a single vertex has no edges. We can therefore apply the stationarity criterion of Razborov [Razborov2007, Corollary 4.6(a)]. At an edge-density maximizer \phi_0, take M_1=\rho, the objective f(t)=-t on \mathbb R, and g=d. The choice f(t)=-t makes \phi_0 a minimizer of the objective in that corollary. It gives \phi_0(\left\llbracket d\cdot\partial_1\rho\right\rrbracket_1)=0, hence \phi_0(S)=0.

Proposition

Proposition 3.1

807P

If Q_b\succeq 0 for every block and the coefficients in (?) satisfy c_G\le B for every G\in\mathcal F^0_m, then \pi(K_4^{(3)})\le B.

Proof of Proposition 3.1

By Proposition 2.4, choose a maximizer \phi with \phi(\rho)=\pi(K_4^{(3)}). By Proposition 2.6, \phi(\tau S)=0 for every \tau\in\mathbb R, while each quadratic-form contribution is nonnegative by Theorem 2.5. Applying \phi to (?) and using Lemma 2.3 gives

\pi(K_4^{(3)})=\phi(\rho) \le\phi\!\left(\rho+ \sum_b\left\llbracket \mathbf F_b^{\mathsf T}Q_b\mathbf F_b\right\rrbracket_{\sigma_b} +\tau S\right) \le\max_{G\in\mathcal F^0_m}c_G\le B.\qedhere
Proposition

Proposition 3.2

TIXV

For the blocks, integer factor vectors, and stationarity multiplier specified above, set

L=5040M^2=322{,}560{,}000{,}000{,}000{,}000, \qquad C_7(G)=L\,c_G,

where c_G is given by (?). Then, C_7(G)\in\mathbb Z for every G\in\mathcal F^0_7, and

\max_{G\in\mathcal F^0_7}C_7(G)=179{,}920{,}412{,}317{,}534{,}044.

Consequently, these data form an exact certificate with

\max_{G\in\mathcal F^0_7}c_G=\frac{14993367693127837}{26880000000000000}=B.
Proof of Proposition 3.2

The factorization of each Q_b is a sum of outer products of rational vectors, so Q_b\succeq 0. For the block sizes (s_b,\ell_b)=(1,4),(3,5),(4,5),(5,6), the average in (?) runs over, respectively, 140, 1260, 5040, and 5040 equally weighted choices of root-label assignments and ordered pairs of disjoint non-root sets. All four numbers divide 5040. The two averages in (?) have respectively 140 and 630 equally likely choices, which also divide 5040. Since the edge-density denominator is 35 and M^2\tau is an integer, (?) implies C_7(G)\in\mathbb Z.

To verify (3.2), we take one representative of each of the 964 tetrahedron-free 3-graphs on six vertices, add one vertex, and consider all 2^{15} choices of edges containing it. Retaining the tetrahedron-free extensions gives a list covering every member of \mathcal F^0_7 up to isomorphism: deleting any vertex from such a 3-graph leaves a member of the six-vertex catalog. Repeated isomorphism classes do not affect the maximum. The exhaustive integer evaluations detailed in Appendix ? establish (3.2) on this list; the counting formulas are given in Appendix ?. Dividing the maximum by L gives B=\frac{14993367693127837}{26880000000000000}. Together with the rational factorizations, this proves that the data form the claimed exact certificate.

Proposition

Proposition 4.1

6HVG

Let G\in\mathcal F^0_7. For each five-vertex set U\subseteq V(G) with G[U]\cong\sigma, choose a labeling \theta_U:[5]\hookrightarrow V(G) with image U that induces \sigma. Let \{x_U,y_U\}=V(G)\setminus U and set i_U=i(\theta_U,x_U) and j_U=i(\theta_U,y_U). Then,

T_\sigma(G;Q_b)=\sum_{\substack{U\subseteq V(G),\ |U|=5\\G[U]\cong\sigma}} (\widehat Q_b)_{i_U,j_U}.

Each summand is independent of the choice of \theta_U and of the ordering of x_U,y_U.

Proof of Proposition 4.1

Fix one such set U. Every labeling with image U that induces \sigma differs from \theta_U by a unique automorphism of \sigma. Hence

\sum_{\substack{\theta([5])=U\\\theta\text{ induces }\sigma}} (Q_b)_{i(\theta,x_U),i(\theta,y_U)} =\sum_{g\in\operatorname{Aut}(\sigma)}(Q_b)_{g\cdot i_U,g\cdot j_U} =(\widehat Q_b)_{i_U,j_U}.

Summing over all such U gives (4.1). Changing \theta_U only permutes the automorphisms in the sum, so the value is independent of the chosen labeling. It is also unchanged when x_U and y_U are interchanged, since \widehat Q_b is symmetric.

Source and licence

The text on this page is from arXiv:2609.27495v2 by Gyeongwon Jeong, Seonghun Park, Seonghyuk Im, Joonkyung Lee and Hongseok Yang, under the licence CC BY 4.0. The corpus made these changes: it read the items out of the LaTeX source, it expanded the macros of the paper, it put each proof after its statement, and it converted the text to HTML. The formulas are the TeX of the paper, and KaTeX renders them. A formula or a reference can look different from the PDF. The page is published under CC BY 4.0.

The record of each item is in analysis/objects/2609/2609.27495.jsonl.