Further improving of upper bound on a geometric Ramsey problem
Eryk Lipka

TL;DR
This paper refines the upper bound on the dimension needed to guarantee a monochromatic coplanar 4-vertex clique in any 2-coloring of the hypercube's edges, improving previous bounds significantly.
Contribution
It improves the known upper bound on the geometric Ramsey problem from a triple exponential tower to a lower triple exponential tower.
Findings
Reduced the upper bound from 2↑↑↑6 to 2↑↑↑5.
Provided a tighter estimate for the minimal dimension n.
Revisited and improved upon previous combinatorial bounds.
Abstract
We consider following geometric Ramsey problem: find the least dimension such that for any 2-coloring of edges of complete graph on the points there exists 4-vertex coplanar monochromatic clique. Problem was first analyzed by Graham and Rothschild and they gave an upper bound: , where . In 2014 Lavrov, Lee and Mackey greatly improved this result by giving upper bound . In this paper we revisit their estimates and reduce upper bound to
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsLimits and Structures in Graph Theory · Advanced Graph Theory Research · Advanced Topology and Set Theory
Further improving of upper bound on a geometric Ramsey problem
Eryk Lipka
Zaremba Society of Mathematicians – Students of the Jagiellonian University
Institute of Mathematics of the Pedagogical University of Cracow
(May 2019)
Abstract
We consider following geometric Ramsey problem: find the least dimension such that for any 2-coloring of edges of complete graph on the points there exists 4-vertex coplanar monochromatic clique. Problem was first analyzed by Graham and Rothschild [1] and they gave an upper bound: , where . In 2014 Lavrov, Lee and Mackey [2] greatly improved this result by giving upper bound . In this paper we revisit their estimates and reduce upper bound to .
1 Setting
Definition 1**.**
Given let Hales-Jewett number be the least integer with the following property. For any -coloring of there exists an injective function such, that
[TABLE]
and is -monochromatic.
Definition 2**.**
Given let Tic-Tac-Toe number be the least integer with the following property. For any -coloring of there exists an injective function such, that
[TABLE]
and is -monochromatic. Image of such a function is called a -dimensional Tic-Tac-Toe Subspace.
Definition 3**.**
Given let be the smallest dimension such that for every edge-coloring of a complete graph on the points there exists an injective function with
[TABLE]
and all edges between the points of have the same color.
In particular, is the smallest integer , such that for every edge-coloring of a complete graph on the points there exist four coplanar vertices such that all six edges between them are monochromatic. Our goal is to give a better upper bound for that value. It has been proven in [2], that , and then, using obvious inequality it was shown that . Our approach is to not use the Hales-Jewett function, because TTT and HJ have similar growth rate, but initial values of TTT are much smaller.
Definition 4**.**
Given let be the least integer with the following property. For any -coloring of there exists -coloring of and an injective function such, that
[TABLE]
[TABLE]
[TABLE]
In other words, values and are not distinguished by induced coloring of .
Lemma 1**.**
Let , then , where
[TABLE]
Proof.
This is straightforward conclusion from chapter 1 of [3]. This fact was used to show, that . ∎
Lemma 2**.**
Let and be defined as above, then
Proof.
For it is obviously true as for . By induction it is true for any because
[TABLE]
∎
Lemma 3**.**
Let , then .
Proof.
First, we notice as a line connecting any two points in has property (1), so we just need to have more points than colors. Define , then by pigeonhole principle . Because then . ∎
Corollary 1**.**
By carefully repeating previous proof we can get even better estimate for certain values, in particular for we have so .
2 Main Result
Lemma 4**.**
For
[TABLE]
Proof.
We will basicaly repeat proof of Lemma 1.4 from [3], but with TTT instead of Hales-Jewett function. Define
[TABLE]
Let be any -coloring of , by definition of Cub we have induced -coloring of and embedding such that properties (2),(3),(4) hold (for ). Again, by definition of Cub we have induced -coloring of and embedding such that properties (2),(3),(4) hold (for ). By definition of TTT there exists an injective function with property (1) and its image is -monochromatic. Let be defined as , also let be natural embeddings.
Define as
[TABLE]
It is easy to check, that is -monochromatic and
[TABLE]
Now, we define function in a following way
[TABLE]
This function satisfies , so , and image of is a Tic-Tac-Toe subspace. Because and have property (4) this image is also -monochromatic, so ∎
Lemma 5**.**
.
Proof.
[TABLE]
∎
Lemma 6**.**
.
Proof.
[TABLE]
∎
Theorem 1**.**
.
Proof.
From [2] we know, that , so . ∎
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] R.L. Graham, B.L. Rothschild, Ramsey’s theorem for n-parameter sets , Trans. Amer. Math. Soc. 159 (1971): 257–292.
- 2[2] M. Lavrov, M. Lee, J. Mackey, Improved upper and lower bounds on a geometric Ramsey problem , European J. Combin. 42 (2014): 135–144.
- 3[3] S. Shelah, Primitive recursive bounds for van der Waerden numbers , J. Amer. Math. Soc. 1.3 (1988): 683-697.
