Alphabeta Math
TheoremStatement: AI-adaptedProof: AI-adaptedPipeline-generatedjudge pass (gpt-5.6-terra)audited 2026-09-14
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Milman converse for compact generating sets

Statement

Let K be a compact convex subset of a locally convex Hausdorff real or complex topological vector space, and let A⊆K. If

K=co⁡‾(A),

then ext⁡K⊆A‾. In particular, if A is compact and generates K in this sense, then ext⁡K⊆A.

Facts & Assumptions

Given: A locally convex Hausdorff real or complex TVS X, a compact convex K⊆X, and A⊆K with K=co⁡‾(A).

[F1]

Extreme points are characterized by strict two-endpoint convex representations (Extreme point and face).

[F2]

Every zero-neighborhood in a locally convex TVS contains an open convex zero-neighborhood (Local convexity, convex and balanced sets, and the continuous dual).

[F5]

The convex hull of finitely many nonempty compact convex sets is compact, is closed in a Hausdorff TVS, and has the displayed one-point-from-each-set representation (Convex closures and hulls of finitely many compact convex sets).

[F6]

If N is a natural number and F is a function with domain N whose values are nonempty, then the family F[N] has a choice function in ZF. Repetitions among the listed values are allowed (Every natural-number-indexed list of nonempty sets has a choice function on its family of values).

Proof

technique · contradiction from a finite compact-convex decomposition
1.1F3F4given

The conclusion is immediate if K=∅. Otherwise A≠∅, because the closed convex hull of the empty set is empty. Put B=A‾. Since K is closed by [F3] and contains A, one has B⊆K; hence B is closed in compact K and compact by [F4].

2.1F2step 1.1givenassume-contra

Assume for contradiction that w∈ext⁡K∖B. The open set X∖B contains w, so translation gives a zero-neighborhood V with w+V⊆X∖B. Continuity of subtraction at (0,0) gives a zero-neighborhood W with W−W⊆V, and [F2] gives an open convex zero-neighborhood U⊆W. Thus (w+U−U)∩B=∅.

3.1F6step 1.1step 2.1

The family {b+U:b∈B} is an open cover of the nonempty compact set B. Compactness supplies a listed subcover O1,…,On with n≥1. For i∈n={0,…,n−1} put Ci={b∈B:b+U=Oi+1}. Each Ci is nonempty because Oi+1 belongs to the displayed cover. Apply [F6] to the function i↦Ci with domain n; if c is the resulting choice function on its family of values, set bi+1=c(Ci). Then bi+1∈B, Oi+1=bi+1+U, and hence B⊆⋃j=1n(bj+U). Put Bj=B∩(bj+U) and Kj=co⁡‾(Bj). Each Bj is nonempty because it contains bj.

4.1F4step 2.1step 3.1

Each Kj is a nonempty compact convex subset of K: the closed convex set K contains Bj, hence its closed convex hull, and Kj is closed in compact K, so [F4] applies. Moreover w∉Kj. Indeed co⁡(Bj)⊆bj+U by convexity; if w lay in its closure, the open neighborhood w+U of w would meet bj+U, giving bj∈w+U−U, contrary to bj∈B and step 2.1.

5.1F5step 3.1step 4.1given

Let H=co⁡(K1∪⋯∪Kn). By [F5], H is compact and therefore closed in the Hausdorff ambient space, and every point of H is ∑j=1ntjxj with xj∈Kj, tj≥0, and ∑jtj=1. Since A⊆B⊆⋃jBj⊆H, closedness and convexity of H give K=co⁡‾(A)⊆H; conversely every Kj⊆K and K is convex, so H⊆K. Hence H=K.

6.1step 4.1step 5.1

Apply the representation in step 5.1 to w: write w=∑jtjxj with xj∈Kj. If exactly one coefficient is positive, it equals one and gives w=xj∈Kj, contradicting step 4.1. Otherwise, for each i with ti>0 one has 0<ti<1 and may write w=tixi+(1−ti)yi, where yi=(1−ti)−1∑j≠itjxj∈K.

7.1F1step 2.1step 4.1step 6.1discharge-contradiction

Since w is extreme, [F1] applied to the strict representation in step 6.1 gives xi=w for every positive coefficient ti. At least one coefficient is positive, so w∈Ki for some i, again contradicting step 4.1. Therefore no such w exists and ext⁡K⊆B=A‾.

8.1F3step 1.1step 7.1discharge-contradiction∎

If A is compact, then it is closed by [F3], so A‾=A and step 7.1 gives ext⁡K⊆A. This proves both assertions, including the empty case from step 1.1.

Depends on

Used by

Dependency tree · two levels

25 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources