Skip to content
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
71 changes: 71 additions & 0 deletions FormalConjectures/OptimizationConstants/1a.lean
Original file line number Diff line number Diff line change
@@ -0,0 +1,71 @@
/-
Copyright 2025 The Formal Conjectures Authors.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
https://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-/

import FormalConjectures.Util.ProblemImports

/-!
# Tao's Optimization constant 1a / An autocorrelation constant related to Sidon sets
*References:*
- [Tao's optimization constant 1a](https://teorth.github.io/optimizationproblems/constants/1a.html)
- [M2010] Matolcsi, Máté, and Carlos Vinuesa. "Improved bounds on the supremum of autoconvolutions."
Journal of mathematical analysis and applications 372.2 (2010): 439-447. [arXiv:0907.1379](https://arxiv.org/abs/0907.1379)
- [Y2026] Yuksekgonul, Mert et al., "Learning to Discover at Test Time," 2026, [arXiv:2601.16175](https://arxiv.org/abs/2601.16175)
-/

open Set

namespace Constant1a

/--
A real number satisfying a certain inequality about (auto)convolutions and $L^2$-norms of functions.
Such numbers are related to the maximal size of Sidon sets in additive combinatorics. -/
def IsL2Bound (C : ℝ) : Prop :=
∀ ⦃f : ℝ → ℝ⦄, 0 ≤ f → C * (∫ x in (- 1 / 4)..(1 / 4), f x) ^ 2
≤ sSup {r | ∃ t ∈ Icc (1 / 2 : ℝ) 1, r = ∫ x, f (t - x) * f x}

/-- **Tao's Optimization constant 1a / An autocorrelation constant related to Sidon sets**:
The biggest real number satisfying a certain inequality about integral. -/
noncomputable def C1a : ℝ := sSup {C : ℝ | IsL2Bound C}

/-- The best known lower bound, proven by Matolcsi-Vinuesa in [M2010]-/
@[category research solved, AMS 05 11 26]
theorem c1a_lower_bound : 1.2748 ≤ C1a := by
sorry

/-- The best known upper bound, proven by Yuksekgonul et al. in [Y2026] -/
@[category research solved, AMS 05 11 26]
theorem c1a_upper_bound : C1a ≤ 1.5029 := by
sorry

/-- How can the upper bound be improved? -/
@[category research open, AMS 05 11 26]
theorem c1a_le : C1a ≤ answer(sorry) := by
Copy link
Collaborator

Choose a reason for hiding this comment

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

Suggested change
theorem c1a_le : C1a ≤ answer(sorry) := by
theorem c1a_le : ∃ (c : ℝ), c < 1.5029C1a ≤ c := by

Might be a version without answer(sorry) or

theorem c1a_le : ∃ (c : ℝ), c = answer(sorry) ∧ c < 1.5029 ∧ C1a ≤ c := by

has the advantage, that no manually check is needed if a good answer was found.
I think it would be worth the effort to think about how to best do this if we are doing this for many optimization problems!

Perhaps we could add some definitions that work analogously.
Pinging also @Paul-Lez and @YaelDillies and @smmercuri and @callesonne and @eric-wieser
for ideas.

Here's an adhoc way I did it at the time: this is https://teorth.github.io/optimizationproblems/constants/1b.html
(probably needs update of the constants, btw)

Find a better lower bound!
-/
@[category research open, AMS 5 11]
theorem erdos_36.variants.lower:
∃ (c : ℝ), 0.379005 < c ∧ c ≤ atTop.liminf MinOverlapQuotient ∧ c = answer(sorry) := by
sorry
/--
Find a better upper bound!
-/
@[category research open, AMS 5 11]
theorem erdos_36.variants.upper :
∃ (c : ℝ), c < 0.380926853433087 ∧ atTop.limsup MinOverlapQuotient ≤ c ∧ c = answer(sorry) := by
sorry
/--
The limit of `MinOverlapQuotient` exists and it is less than $0.385694$.
-/
@[category research solved, AMS 5 11]
theorem erdos_36.variants.exists : ∃ c, atTop.Tendsto MinOverlapQuotient (𝓝 c) ∧ c < 0.385694 := by
sorry
/--
Find the value of the limit of `MinOverlapQuotient`!
-/
@[category research open, AMS 5 11]
theorem erdos_36 : atTop.Tendsto MinOverlapQuotient (𝓝 answer(sorry)) := by
sorry

This is also a thing to consider: we might want to introduce a mechanism to pull the constants from
https://teorth.github.io/optimizationproblems/constants
So perhaps they better go into an attribute?

Copy link
Member

Choose a reason for hiding this comment

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

I can have a go at implementing an attribute for those.

sorry

/-- How can the lower bound be improved? -/
@[category research open, AMS 05 11 26]
theorem c1a_ge : answer(sorry) ≤ C1a := by
sorry

/-- What is the exact value of the constant? -/
@[category research open, AMS 05 11 26]
theorem c1a_eq : C1a = answer(sorry) := by
sorry

-- TODO : Formalise relationship with Sidon sets.

end Constant1a
Loading