Skip to content

Conversation

@krrish175-byte
Copy link

Closes #1737

This PR formalizes Green's Open Problem 82, which asks: for a finite set A ⊂ ℤ of size n, what is the minimum number of zeros that ∑_{a∈A} cos(2πaθ) must have on ℝ/ℤ?

Signed-off-by: krrish175-byte <krrishbiswas175@gmail.com>
@github-actions github-actions bot added the green-problems Problems from https://people.maths.ox.ac.uk/greenbj/papers/open-problems.pdf label Jan 26, 2026
$A \subset \mathbb{Z}$ with $|A| = n$.
-/
@[category research open, AMS 11 42]
theorem green_82 (n : ℕ) (hn : 1 ≤ n) :
Copy link
Contributor

Choose a reason for hiding this comment

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

n probably shouldnt be in the context as usual

Comment on lines 47 to 48
⨅ (A : Finset ℤ) (_ : A.card = n),
({θ : ℝ | θ ∈ Ico 0 1 ∧ ∑ a ∈ A, cos (2 * π * a * θ) = 0} : Set ℝ).ncard =
Copy link
Contributor

Choose a reason for hiding this comment

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

Better write this as an inequality and not take sup?

@felixpernegger
Copy link
Contributor

The green problem list mentions known bounds, please include them.

@[category research open, AMS 11 42]
theorem green_82 (n : ℕ) (hn : 1 ≤ n) :
⨅ (A : Finset ℤ) (_ : A.card = n),
({θ : ℝ | θ ∈ Ico 0 1 ∧ ∑ a ∈ A, cos (2 * π * a * θ) = 0} : Set ℝ).ncard =
Copy link
Contributor

Choose a reason for hiding this comment

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

This is not what the problem is asking. Why do you have to limit $\theta$ to Ico 0 1?

Signed-off-by: krrish175-byte <krrishbiswas175@gmail.com>
@krrish175-byte
Copy link
Author

Thanks for the review. Sorry, I missed those. I have updated the code as you asked for:

  • Moved n inside the theorem statement.
  • Reformulated the statement using an inequality.
  • Added the known lower bound $\Omega(\log n / \log \log n)$.

Regarding Ico 0 1: Since the trigonometric polynomial is 1-periodic, counting zeros in the fundamental domain [0, 1) seemed like the most direct way to represent the count on R / Z. Let me know if you would prefer using UnitAddCircle instead. Thanks for the review again!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

green-problems Problems from https://people.maths.ox.ac.uk/greenbj/papers/open-problems.pdf

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Green's Open Problems #82

2 participants