Skip to content

feat(ErdosProblems): 15x ramsey theory formalizations#3588

Open
ryantuck wants to merge 6 commits intogoogle-deepmind:mainfrom
ryantuck:erdos-ramsey
Open

feat(ErdosProblems): 15x ramsey theory formalizations#3588
ryantuck wants to merge 6 commits intogoogle-deepmind:mainfrom
ryantuck:erdos-ramsey

Conversation

@ryantuck
Copy link
Contributor

@mo271
Copy link
Collaborator

mo271 commented Mar 17, 2026

Thanks!

Some first remarks:

  • It makes sense to group them like this!
  • It would be good to move re-usable definitions to FormalConjecturesForMathlib and then use them in the files.
  • feat(Wikipedia): add Ramsey numbers (R(5,5)) #2436 was just merged adding a definition of Ramsey number (including some notation). It would be good to move what in that file to appropriate files in FormalConjecturesForMathlib and use it from there.
  • Note that we already have some Ramsey related concepts, like hypergraphRamsey and sizeRamsey and antiRamseyNum (didn't look in detail if any of this can be reused in one of the 15 problems at hand.

@ryantuck
Copy link
Contributor Author

@mo271 thanks for the feedback!

As advised, I had Opus 4.6 do the following (fine-grain markdown session logs included to show the work):

1 - 3c168fa consolidated reused definitions among my erdos contributions (session logs)

"from these 15 problems, i want to abstract out reused definitions and include them in the appropriate place within FormalConjecturesForMathlib. identify code reuse and refactor accordingly."

2 - 6b288cd also consolidated wikipedia Ramsey code for reuse (session logs)

"FormalConjectures/Wikipedia/RamseyNumbers.lean was added recently. further consolidate common code usage, deferring to the implementation from the wikipedia lean file if a choice is needed."

3 - 45fd907 holistically assessed whether all Ramsey-related code was in an appropriate place or needed further consolidation, and clarified some references in the erdos file implementations (session logs)

"there are multiple ramsey-related files in FormalConjecturesForMathlib. Assess whether it makes sense to keep them separate or to consolidate them and then attend to the relevant files within the FormalConjectures dir."

Copy link
Member

@YaelDillies YaelDillies left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please extend my comments accordingly to the other problems

Comment on lines +48 to +50
∀ ε : ℝ, ε > 0
∃ L₀ : ℕ, ∀ l : ℕ, l ≥ L₀ →
|(graphRamseyNumber k (l + 1) : ℝ) / (graphRamseyNumber k l : ℝ) - 1| ≤ ε := by
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you state this using Filter.Tendsto, Filter.atTop and TopologicalSpace.nhds instead?

Comment on lines +36 to +47
/--
Erdős Problem 1014 [Er71, p.99]:
For fixed $k \geq 3$,
$$\lim_{l \to \infty} R(k, l+1) / R(k, l) = 1,$$
where $R(k, l)$ is the Ramsey number.
Formulated as: for every $\varepsilon > 0$, there exists $L_0$ such that for all $l \geq L_0$,
$|R(k, l+1) / R(k, l) - 1| \leq \varepsilon$.
-/
@[category research open, AMS 5]
theorem erdos_1014 (k : ℕ) (hk : k ≥ 3) :
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you state the k = 3 separately since it already is open?

*Reference:* [erdosproblems.com/1014](https://www.erdosproblems.com/1014)
[Er71] Erdős, P., _Topics in combinatorial analysis_, pp. 95-99, 1971.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The article seems to only be 19 pages long. Please check all references!

Comment on lines +58 to +59
∃ K₀ : ℕ, ∀ k : ℕ, k ≥ K₀ →
(diagRamseyNumber k : ℝ) ≥ C * (k : ℝ) * (2 : ℝ) ^ ((k : ℝ) / 2) := by
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Again, please use Filter primitives here


*Reference:* [erdosproblems.com/1030](https://www.erdosproblems.com/1030)

If $R(k,l)$ is the Ramsey number then prove the existence of some $c > 0$ such that
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The only Ramsey number?? 😁

See also problems [544](https://www.erdosproblems.com/544) and
[1014](https://www.erdosproblems.com/1014).

OEIS: [A000791](https://oeis.org/A000791), [A059442](https://oeis.org/A059442).
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

What's the relevance of the first one?

@YaelDillies YaelDillies added the awaiting-author The author should answer a question or perform changes. Reply when done. label Mar 20, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

awaiting-author The author should answer a question or perform changes. Reply when done. erdos-problems Erdős Problems wikipedia

Projects

None yet

3 participants