Monotone continuity characterizes convex µ-compact sets
OpenQAlgorithms.Papers.Independent.Support50.P33.monotoneContinuityCharacterizationLet be a real Hausdorff locally convex topological vector space, and let be convex and von Neumann bounded. Assume the relative topology of is generated by a countable subfamily of the ambient defining seminorms, and that is complete and separable in the resulting metrizable additive uniformity. The set is allowed to be empty. All regularity and convergence below are relative to .
Write for the Borel probability measures on with the weak convergence topology. The barycenter is the Pettis integral of inclusion: every continuous real linear functional is integrable against , and . The set is -compact when the barycenter preimage of every compact subset of is compact in ; it is pointwise -compact when every barycenter fiber is compact.
Write for the bounded continuous real functions on . For an extended-valued cost , let be its greatest convex minorant and its greatest lower semicontinuous convex minorant. A finite lower bound means for all for some , so positive infinity is permitted but negative infinity is excluded on .
The following three properties are equivalent:
- is -compact.
- Every pointwise nondecreasing sequence in with pointwise limit satisfies
- Every pointwise nondecreasing sequence of lower semicontinuous functions , each with its own finite lower bound, with pointwise limit , satisfies the same displayed identity, with convergence in the extended real line. The limit need not be finite or bounded.
When these equivalent properties hold, the displayed identity also holds for every pointwise nonincreasing sequence of bounded real lower semicontinuous functions whose pointwise limit is bounded real and lower semicontinuous. Boundedness is required individually, without a common two-sided bound across the sequence. The notation here denotes the separate limit, not a designated first term.
This characterizes compactness of barycenter preimages by sequential continuity of convex closure, and includes the decreasing-sequence consequence.
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P33
noncomputable section
open Set Filter MeasureTheory TopologicalSpace
open scoped Topology ENNReal
namespace QAlgorithms.Papers.Independent.Support50.P33
open QAlgorithms.Papers.Independent.Frontier18.P10
open QAlgorithms.Papers.Independent.Dependency50.P25
attribute [local instance] QAlgorithms.Papers.Independent.Support50.P33.paperMeasurable
attribute [local instance] QAlgorithms.Papers.Independent.Support50.P33.paperBorel
section LocallyConvexResults
variable {E : Type*} [UniformSpace E] [AddCommGroup E] [Module ℝ E]
[hgroup : IsUniformAddGroup E] [hsmul : ContinuousSMul ℝ E]
[hlc : LocallyConvexSpace ℝ E] [hhaus : T2Space E]
include hgroup hsmul hlc hhaus
/-- Theorem, arXiv:1004.3792v1, PDF p.3, including the final decreasing
sequence assertion. The two equivalences encode (i) iff (ii) iff (iii).
All envelopes in this theorem are closed convex envelopes. In (iii) every
term has a finite real lower bound, and positive infinite costs and limits
are allowed. The separate limit is not the first sequence term. -/
theorem monotoneContinuityCharacterization (A : Set E)
(hA : GeneralBarycenterSpace A) (hc : Convex ℝ A) :
(MuCompact A ↔
∀ (f : ℕ → E → ℝ) (fLimit : E → ℝ),
(∀ n, BoundedContinuousOn A (f n)) → BoundedContinuousOn A fLimit →
(∀ x ∈ A, Monotone (fun n => f n x)) →
(∀ x ∈ A, Tendsto (fun n => f n x) atTop (𝓝 (fLimit x))) →
∀ x ∈ A, Tendsto
(fun n => closedConvexHullFunction A (fun y => (f n y : EReal)) x)
atTop (𝓝 (closedConvexHullFunction A (fun y => (fLimit y : EReal)) x))) ∧
(MuCompact A ↔
∀ (f : ℕ → E → EReal) (fLimit : E → EReal),
(∀ n, LowerSemicontinuous (fun x : A => f n x)) →
(∀ n, ∃ c : ℝ, ∀ x ∈ A, (c : EReal) ≤ f n x) →
(∀ x ∈ A, Monotone (fun n => f n x)) →
(∀ x ∈ A, Tendsto (fun n => f n x) atTop (𝓝 (fLimit x))) →
∀ x ∈ A, Tendsto (fun n => closedConvexHullFunction A (f n) x)
atTop (𝓝 (closedConvexHullFunction A fLimit x))) ∧
(MuCompact A →
∀ (f : ℕ → E → ℝ) (fLimit : E → ℝ),
(∀ n, LowerSemicontinuous (fun x : A => f n x)) →
(∀ n, ∃ M : ℝ, ∀ x ∈ A, |f n x| ≤ M) →
LowerSemicontinuous (fun x : A => fLimit x) →
(∃ M : ℝ, ∀ x ∈ A, |fLimit x| ≤ M) →
(∀ x ∈ A, Antitone (fun n => f n x)) →
(∀ x ∈ A, Tendsto (fun n => f n x) atTop (𝓝 (fLimit x))) →
∀ x ∈ A, Tendsto
(fun n => closedConvexHullFunction A (fun y => (f n y : EReal)) x)
atTop (𝓝 (closedConvexHullFunction A (fun y => (fLimit y : EReal)) x))) := by sorry
end LocallyConvexResults
end QAlgorithms.Papers.Independent.Support50.P33
end
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
Let be the ambient real vector space with its uniform structure and induced topology, and let be closed, convex, and von Neumann bounded, with complete, separable, and metrizable. For a Borel probability measure on , define by choosing a point such that every continuous real linear functional is integrable on against and satisfies , if such a point exists, and setting otherwise. Write for the assertion that is compact in the space of probability measures on for every compact . For , define as the supremum of over all whose restriction to is lower semicontinuous, which satisfy on , and for which whenever , , and ; these operations use extended-real arithmetic, including zero coefficients. The theorem asserts the conjunction of three statements: (1) holds if and only if, for every sequence , indexed by all , and every , if every and is continuous on and bounded there in absolute value by a real constant, is nondecreasing in for every , and for every , then for every . (2) holds if and only if the same envelope-convergence conclusion holds for every sequence and every such that each is lower semicontinuous, each has a real lower bound for all , and increases nondecreasingly to at every ; no separate regularity assumption on is imposed. (3) If holds, then for every whenever are lower semicontinuous on , each is bounded there in absolute value by a real constant, and decreases nonincreasingly to at every . All bounds on individual sequence terms may depend on , and all envelope convergence is in the extended-real topology. The set may be empty, in which case assertions quantified over are vacuous.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.