OAI.Problem358.FixedMassIntercepts.exists_actual_atomic_intercept_path: intermediate lemma from OpenAI, Sharp One-Third Stability of Brenier Maps
ProvedOAI.Problem358.FixedMassIntercepts.exists_actual_atomic_intercept_pathAn intermediate lemma of the split of OpenAI's proof of OAI.Problem358.one_third_stability, carved from OpenAI's openai/math development for this platform. The statement and its proof are OpenAI's (OpenAI, Sharp One-Third Stability of Brenier Maps, OpenAI Math Release, September 25, 2026, result family 374, github.com/openai/math, Apache License 2.0); in Lean it is the declaration OAI.Problem358.FixedMassIntercepts.exists_actual_atomic_intercept_path of lean/OAI/Analysis/Brenier/Brenier/CanonicalPaths.lean at commit adc7f12, ported to Lean 4.33.1 and Mathlib 0df444a.
The statement, as published (namespace OAI.Problem358.FixedMassIntercepts, with open Set MeasureTheory, open scoped BigOperators NNReal ContDiff):
theorem exists_actual_atomic_intercept_path
{d : ℕ} {ι : Type*} [Fintype ι] [Nonempty ι]
[MeasurableSpace ι] [MeasurableSingletonClass ι]
{K : Set (E d)} (hK : IsCompact K) (hconv : Convex ℝ K)
(hKi : (interior K).Nonempty)
(m : ι → ℝ) (hm : ∀ i, 0 < m i) (hsum : ∑ i, m i = 1)
(y z : ι → E d) (R : ℝ≥0)
(hbounded : ∀ᵐ x ∂uniformMeasure K, ‖x‖ ≤ R)
(hinj : ∀ t ∈ Ioo (0 : ℝ) 1,
Function.Injective (fun i => (1-t) • y i + t • z i)) :
∃ c r : ι → ℝ → ℝ,
(∀ i, Measurable (c i)) ∧
(∀ i, ContinuousOn (r i) (Ioo 0 1)) ∧
(∀ i t, t ∈ Ioo 0 1 → HasDerivAt (c i) (r i t) t) ∧
(∀ t ∈ Ioo 0 1, ∀ i,
(uniformMeasure K).real (CellGeometry.closedCell K
(fun j => (1-t) • y j + t • z j) (fun j => c j t) i) = m i) ∧
(∀ t ∈ Ioo 0 1, (∑ i, m i * c i t) = 0) ∧
∀ t ∈ Ioo 0 1, (∑ i, m i * r i t ^ 2) ≤
162 * (R : ℝ) ^ 2 * ∑ i, m i * ‖y i-z i‖ ^ 2
Definitions. The statement is written with OAI.Problem358.CellGeometry.closedCell (definitions bundle OAIQ5T2000 of this split); OAI.Problem358.E and OAI.Problem358.uniformMeasure (the published bundle Def_Brenier).
Used by. The proof of OAI.Problem358.one_third_stability of this split imports it.
import Mathlib import Definitions.Def_Brenier import Definitions.Def_OAIQ5T2000
section
set_option linter.all false
set_option linter.unusedVariables false
set_option linter.unusedTactic false
set_option linter.unreachableTactic false
namespace OAI
noncomputable section
open Set MeasureTheory
open scoped BigOperators NNReal ContDiff
namespace Problem358.FixedMassIntercepts
theorem exists_actual_atomic_intercept_path
{d : ℕ} {ι : Type*} [Fintype ι] [Nonempty ι]
[MeasurableSpace ι] [MeasurableSingletonClass ι]
{K : Set (E d)} (hK : IsCompact K) (hconv : Convex ℝ K)
(hKi : (interior K).Nonempty)
(m : ι → ℝ) (hm : ∀ i, 0 < m i) (hsum : ∑ i, m i = 1)
(y z : ι → E d) (R : ℝ≥0)
(hbounded : ∀ᵐ x ∂uniformMeasure K, ‖x‖ ≤ R)
(hinj : ∀ t ∈ Ioo (0 : ℝ) 1,
Function.Injective (fun i => (1-t) • y i + t • z i)) :
∃ c r : ι → ℝ → ℝ,
(∀ i, Measurable (c i)) ∧
(∀ i, ContinuousOn (r i) (Ioo 0 1)) ∧
(∀ i t, t ∈ Ioo 0 1 → HasDerivAt (c i) (r i t) t) ∧
(∀ t ∈ Ioo 0 1, ∀ i,
(uniformMeasure K).real (CellGeometry.closedCell K
(fun j => (1-t) • y j + t • z j) (fun j => c j t) i) = m i) ∧
(∀ t ∈ Ioo 0 1, (∑ i, m i * c i t) = 0) ∧
∀ t ∈ Ioo 0 1, (∑ i, m i * r i t ^ 2) ≤
162 * (R : ℝ) ^ 2 * ∑ i, m i * ‖y i-z i‖ ^ 2 := by
sorry
end Problem358.FixedMassIntercepts
end
end OAI
end