Alphabeta Math
LemmaStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedjudge pass (gpt-6-luna)verified 2026-10-22 (gpt-6.1-sol)
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.

Coarea identity for smooth scalar functions

Statement

Assume Countable Choice. Let Ω⊆Rn be open and let u∈C∞(Ω;R)∩L1(Ω) satisfy ∫Ω∣∇u∣ dx<∞. For t>0 set Et+:={x∈Ω:u(x)>t} and for t<0 set Et−:={x∈Ω:u(x)<t}. Then the level sets have finite perimeter relative to Ω for almost every applicable t, and ∫Ω∣∇u∣ dx=∫0∞P(Et+;Ω) dt+∫−∞0P(Et−;Ω) dt.

Facts & Assumptions

Given: Countable Choice, an open Ω⊆Rn, and a smooth real function u∈L1(Ω) with ∇u∈L1(Ω;Rn).

[A1]

Countable Choice is assumed for the countable chart and surface-measure constructions. (The Axiom of Countable Choice (ACω))

[F1]

The critical values of a smooth scalar function on an open Euclidean set have Lebesgue measure zero. (Morse-Sard for Euclidean maps)

[F2]

A C1 map with nonsingular derivative has a local C1 inverse, and an injective such map on an open set obeys the compactly supported change-of-variables formula. (The Euclidean inverse function theorem, A compactly supported Riemann integrand admits the global change-of-variables formula from a diffeomorphism near the relevant compact preimage)

[F3]

On an m-dimensional graph, Hausdorff measure divided by the repository normalization constant cm has density 1+∣Dh∣2; for C1 graphs this agrees with the chart surface integral. Graph surface measures are Radon in their relative topology, and pushforward to the open ambient domain is Radon when the graph is relatively closed and the total measure finite. (Lipschitz graph area and calculus under Countable Choice, Chart and partition independence of surface measure)

[F4]

For a C1 graph boundary and a compactly supported C1 vector field, integrating its divergence over the subgraph gives the graph flux formula. (The local graph flux calculation)

[F5]

Rational points are dense in Euclidean space and form a countable family, so rational balls give a countable base. (The rationals embed densely in the reals, Q is countably infinite, A product of two at most countable sets is at most countable)

[F6]

A countable open cover of an open Euclidean set admits a locally finite smooth partition of unity subordinate to that cover. (Locally finite smooth partitions of unity on Euclidean open sets)

[F7]

A continuous unit-vector density against a finite Radon surface measure has partition variation equal to that surface measure. (A unit vector density has scalar variation measure)

[F8]

Tonelli permits summing nonnegative chart integrals and interchanging the level and spatial integrations. (Tonelli's theorem for nonnegative measurable functions on a sigma-finite product)

Proof

Given: The hypotheses listed above.

1.1A1F2F3F5F6givenconstruct

Put G:={x∈Ω:∇u(x)≠0}. The balls with rational centers and radii whose closures lie in Ω and on which some ∂iu never vanishes form a countable open cover of G by density and countability [F5]; assign each ball the least such coordinate i. Each ball is convex. The map Ti(x)=(x1,…,xi−1,u(x),xi+1,…,xn) has nonzero Jacobian there; it is injective because along each line parallel to the ith axis, u is strictly monotone. Thus [F2] makes it a C1 diffeomorphism onto its open image. By [F6], choose a locally finite smooth partition of unity (ηj) on G subordinate to these chart balls. On a ball assigned coordinate i, write y=xi^ and t=u(x); the inverse chart has graph form xi=h(y,t), with ∣det⁡DTi∣=∣∂iu∣ and 1+∣Dyh∣2=∣∇u∣∣∂iu∣ by implicit differentiation. Change of variables [F2] and the surface density [F3], applied to the compactly supported nonnegative weight ηj, give ∫Gηj∣∇u∣ dx=∫R∫G∩{u=t}ηj dHn−1 dt.

2.1F1F8step 1.1given

Sum the chart identities of step 1.1 over j. Since ∑jηj=1 on G, Tonelli [F8] gives ∫Ω∣∇u∣ dx=∫RHn−1(G∩{u=t}) dt, because ∇u=0 on Ω∖G. Sard's theorem [F1] says that outside a null set every point of a level is regular, so the integrand agrees almost everywhere in t with Hn−1(Ω∩{u=t}). This also establishes measurability of the level-area function as a countable sum of the chart integrals in step 1.1.

3.1A1F3F4F7step 2.1given∎

Fix a nonzero regular value t for which the area in step 2.1 is finite; throughout the proof Hn−1 denotes the normalized graph measure from [F3], with counting measure if n=1. In that case every graph chart is a single point at fixed level, its empty-coordinate Jacobian is one, and the same substitution identity in step 1.1 is the ordinary one-dimensional substitution formula. Then Γt:=Ω∩{u=t} is a smooth hypersurface. For a compactly supported smooth test field ϕ, its intersection with Γt is compact; local finiteness gives a neighborhood meeting only finitely many chart supports, whose partition sum is one there. Decompose ϕ into the finite chart-supported sum and a remainder vanishing near Γt. The remainder has zero divergence integral over either level set because its restriction to that open set extends as a compactly supported smooth field. Each chart term obeys the graph flux formula [F4]. Summing shows that DχEt+=+(∇u/∣∇u∣)Hn−1∣Γt and DχEt−=−(∇u/∣∇u∣)Hn−1∣Γt. The level Γt is relatively closed in Ω; hence this measure is finite and Radon by the graph area construction [F3], and [F7] gives its partition variation as exactly Hn−1∣Γt. Hence the indicator is BV and its perimeter equals that area. For t>0 and t<0, respectively, Markov's inequality gives ∣Et+∣≤t−1∥u∥1 and ∣Et−∣≤∣t∣−1∥u∥1, so the indicators are in L1. The area integral is finite, so almost every applicable nonzero level is regular and has finite area by step 2.1; the required perimeter equality and almost-everywhere finiteness follow.

Depends on

Used by

Dependency tree · two levels

97 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