Alphabeta Math
LemmaStatement: AI-generatedProof: AI-generated
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.

Smooth positive inversion and composition in the weighted multiplier class

Statement

Fix the exterior and chart data of the weighted Hölder norms, an integer k≥0 and 0<α<1. Write Aβ=Cβk,α(Σ1), with finite arrays given the sum of component norms. All domains below use the full finite-norm Hölder spaces, including nonsmooth elements.

We use the following normed-space meaning of C1. For an open subset U of a real normed vector space X, a map F:U→Y is continuously Fréchet differentiable when, at every u∈U, there is a linear map DF(u):X→Y with finite operator norm ∥L∥op=sup⁡∥h∥X≤1∥Lh∥Y, such that

lim⁡0≠h→0∥F(u+h)−F(u)−DF(u)h∥Y∥h∥X=0,

and ∥DF(v)−DF(u)∥op→0 as v→u. Completeness is not part of this definition.

The set P={a∈A0:inf⁡Σ1a>0} is open. If a∈P, in particular if a smooth a∈A0 satisfies a≥c>0, then a−1∈A0 and multiplication by either a or a−1 is bounded on every Aβ, β∈R. Inversion I:P→A0 is C1, with DI(a)h=−a−2h. The maps Mβ:A0×Aβ→Aβ, Mβ(a,v)=av, and Qβ:P×Aβ→Aβ, Qβ(a,v)=a−1v, are C1, with

DMβ(a,v)[h,w]=hv+aw,DQβ(a,v)[h,w]=−a−2hv+a−1w.

More generally let m,n≥1, let V⊆Rm be open, and let G:V→Rn be smooth. The open domain for composition is

UV={u∈(A0)m:u(Σ1)⊆V, dist⁡(u(Σ1),Rm∖V)>0},

where distance to the empty set is +∞. Then NG:UV→(A0)n, NG(u)=G∘u, is C1, with DNG(u)h=DG(u)h pointwise. A slightly larger attained range means a fixed positive-radius tube about the closure of u(Σ1) contained in V. Bounds for composition use the norm of u and uniform bounds for the derivatives of G through order k+1 on such a tube; continuous differentiability uses derivatives through order k+2 there. These bounds and the derivative-continuity modulus are uniform over the scaled exterior annuli. Thus smooth bounded weighted arguments have the asserted bounded compositions and positive inverse multipliers, while differentiability is asserted on the specified open Hölder domains.

Facts & Assumptions

[F1]

Weighted multiplication satisfies ∥fg∥Aγ+η≤C∥f∥Aγ∥g∥Aη. The supplier's Proof 1.2, 1.3 and 2.1 also gives the Hölder product difference inequality and control of lower-order Hölder seminorms on the fixed scaled annulus and the fixed core chart subsets, even when the latter are not convex. Weighted differentiation and multiplication estimates

[F2]

The zero-weight norm controls the global value supremum and the full order-k norms in every scaled annulus and fixed core chart. It defines a normed vector space; finite arrays use component sums. Weighted Hölder norms on a three-dimensional exterior

[F3]

Continuous coordinate partial derivatives imply total differentiability on chart interiors. The total chain rule, the scalar product and quotient rules on coordinate lines, and permutation of continuous mixed partials give finite iterated chain-rule expansions. Identities at the boundary follow by continuity from the interior. If all partial derivatives exist on a neighbourhood and are continuous at a point, then the map is totally differentiable there with Jacobian derivative The chain rule for total derivatives: D(g∘ f)(a)=Dg(f(a))∘ Df(a) Sums, scalar multiples, products and quotients: (f+g)'(c) = f'(c) + g'(c), (α f)'(c) = α f'(c), (fg)'(c) = f'(c)g(c) + f(c)g'(c), and (f/g)'(c) = (f'(c)g(c) - f(c)g'(c))/g(c)² when g(c) ≠ 0 Continuous mixed partials of order k are invariant under permutations

[F4]

For a scalar C1 function f on an open interval containing [0,1], Taylor's integral formula at order zero reads f(1)−f(0)=∫01f′(t) dt. We apply it componentwise to smooth finite-dimensional functions along segments lying inside V, using the total chain rule for their derivatives. Taylor's theorem with an integral remainder The chain rule for total derivatives: D(g∘ f)(a)=Dg(f(a))∘ Df(a)

[F5]

Closed bounded subsets of finite-dimensional Euclidean space are compact, and a continuous map on a compact metric space is uniformly continuous. Continuous derivatives on our compact tubes are also bounded: continuity supplies a neighborhood bound at each point and a finite subcover supplies one bound for the tube. Heine-Borel in ℝⁿ: with the Euclidean metric a subset of ℝⁿ is compact if and only if it is closed and bounded, and the proof by bisection uses no choice principle; the same holds on the real line Heine-Cantor: a continuous map from a compact metric space to any metric space is uniformly continuous

Proof

Given: The fixed weighted norms, the functions and the finite-dimensional smooth map in the Statement.

Goal: Establish the multiplier and composition bounds, the open domains, and all the stated normed-space derivatives and their continuity.

1.1F2F5givenalgebra

By [F2], ∥u∥(A0)m controls sup⁡∣u∣. Thus a sufficiently small norm perturbation of a point of P preserves a positive lower bound, and a sufficiently small norm perturbation of a point of UV preserves a positive distance from the complement of V. This proves openness in the full spaces. Fix u∈UV and choose δ>0 with 4δ<dist⁡(u(Σ1),Rm∖V), choosing any finite such δ if the complement is empty. The closure S=u(Σ1)‾ is bounded, and its closed 2δ-tube Ku={z:dist⁡(z,S)≤2δ} is closed, bounded and contained in V. Closedness follows from ∣dist⁡(z,S)−dist⁡(z′,S)∣≤∣z−z′∣, which follows directly by taking infima in the triangle inequality. Hence Ku is compact by [F5]. All derivatives of G needed below are bounded and uniformly continuous on Ku. Norm-small h have sup⁡∣h∣<δ/2, so the segments u(p)+th(p) lie in this same tube.

2.1F1F2F3F4step 1.1algebra

Work on the fixed annulus in scaled coordinates, or on a fixed core chart subset, and denote that coordinate set by E. The scaled order-k norms of u and h are bounded uniformly by their zero-weight norms. For a smooth coefficient H on the tube, the segment formula in [F4] gives ∣H(u(p))−H(u(q))∣≤C∣u(p)−u(q)∣ when ∣u(p)−u(q)∣<δ; this segment stays within the tube. For the remaining pairs, boundedness of H gives ∣H(u(p))−H(u(q))∣≤2sup⁡Ku∣H∣ ∣u(p)−u(q)∣/δ. Therefore its Hölder seminorm is bounded by a constant times [u]α;E. For 1≤∣J∣≤k, repeated coordinate chain and product rules express ∂J(G(u)) as a finite sum of a component of DjG(u) times components of derivatives of u of positive orders adding to ∣J∣, with 1≤j≤∣J∣. Apply the coefficient estimate, [F1]'s lower-order seminorm bounds, and its product difference inequality to every factor. This proves the order-k composition bound on each E, using derivatives of G only through order k+1. For k=0 just the coefficient estimate and the value supremum suffice. Taking the annular supremum and adding the core norms proves G(u)∈(A0)n. The same calculation on compact coordinate neighborhoods proves the required local Hölder regularity, and continuity extends the derivative formulas to the boundary.

3.1F1F2F3F4F5step 1.1step 2.1algebra

We next prove continuity of u↦DG(u) in the order-k zero-weight norm, rather than assume it from a formal chain rule. Let H be any component of Dj+1G, 0≤j≤k, and put b=H(u+h)−H(u) and ε=∥h∥(A0)m. Its supremum is at most Cε by [F4]. For a pair with ∣u(p)−u(q)∣<δ, split the increment of b by first replacing h(p) by h(q). That change is at most C∣h(p)−h(q)∣. The remaining difference is the difference of two increments of H along the parallel segments from u(q) to u(p) and from u(q)+h(q) to u(p)+h(q). Apply [F4] on those segments. Uniform continuity of DH on Ku bounds their difference by ω(Cε)∣u(p)−u(q)∣, for a common modulus ω(s)→0 as s→0. All these segments lie in Ku. For a pair with ∣u(p)−u(q)∣≥δ, use the supremum bound on b and ∣u(p)−u(q)∣≤[u]α;E∣p−q∣α. Consequently [b]α;E≤C[h]α;E+(ω(Cε)+Cε/δ)[u]α;E. The right side tends to zero uniformly over all E, by [F1] and [F2]. Derivatives of H used here have order at most k+2 as derivatives of G. In the finite chain-rule expansions for the derivatives through order k of DG(u+h)−DG(u), the coefficient differences therefore tend to zero in C0,α(E), while differences of the derivative factors of u are derivatives of h, whose relevant supremum and Hölder bounds tend to zero by [F1]. Subtract products one factor at a time and use the Hölder product difference inequality. This proves ∥DG(u+h)−DG(u)∥(A0)n×m→0, uniformly over annuli. The proof includes k=0, where there are no differentiated factors of u.

4.1F1F2F3F4F5step 1.1step 2.1step 3.1algebra

Apply [F4] componentwise to t↦G(u(p)+th(p)); smoothness and step 1.1 give its C1 extension to an open interval containing [0,1]. Subtracting its linear part gives the pointwise identity G(u+h)−G(u)−DG(u)h=∫01(DG(u+th)−DG(u))h dt. The integral here is an ordinary finite-dimensional integral at each point, not an assumed integral in a complete function space. Coordinate derivatives through order k may pass under it: for the first derivative the difference quotient equals the integral of the difference quotients, and the segment formula in the coordinate direction and continuity of that derivative on the compact parameter interval imply uniform convergence in t; repeat for the subsequent derivatives. This argument is local on chart interiors; continuous derivative extensions give boundary values. The triangle inequality for finite Riemann sums and their limits, applied also to each two-point difference, gives the integral bound in every chart norm. Using [F1] and step 3.1 thus yields ∥G(u+h)−G(u)−DG(u)h∥(A0)n≤C∥h∥(A0)msup⁡0≤t≤1∥DG(u+th)−DG(u)∥(A0)n×m=o(∥h∥(A0)m). The supremum tends to zero because step 3.1 bounds it by a common modulus of ∥h∥. Step 2.1 applied to DG gives a bounded coefficient multiplier, so h↦DG(u)h is bounded linear by [F1]. That same product bound and step 3.1 give continuity in operator norm. This verifies precisely the normed-space definition in the Statement.

5.1F1F2F3step 1.1step 2.1step 4.1algebra

Take V=(0,∞) and G(s)=s−1. Its domain UV is exactly P, and a positive lower bound and a finite value supremum bound all derivatives of G on a slightly larger positive range. Steps 2.1 and 4.1 therefore give a−1∈A0 and DI(a)h=−a−2h. By [F1], a and a−1 act boundedly on every Aβ. These maps are inverse multiplication operators, since their pointwise products equal one. This includes the original smooth-input claim without restricting the perturbations to smooth functions.

6.1F1step 3.1step 5.1algebra∎

For multiplication, (a+h)(v+w)−av−hv−aw=hw, whose Aβ norm is at most C∥h∥A0∥w∥Aβ by [F1]. This is a quadratic remainder in the sum norm, and the displayed derivative of Mβ varies continuously in operator norm by the same estimate. Write I(a+h)=I(a)−a−2h+r(h), with ∥r(h)∥A0=o(∥h∥A0) from step 5.1. Expanding I(a+h)(v+w) leaves, after subtracting the displayed derivative of Qβ, the terms r(h)v+(I(a+h)−I(a))w, which are o(∥h∥A0+∥w∥Aβ) by [F1] and continuity of I. The coefficient fields I(a) and −a−2 vary continuously in A0 by steps 3.1 and 5.1, so the derivative of Qβ is continuous in operator norm. All claimed maps and bounds follow.

Depends on

Used by · 4

Dependency tree · 2 levels

Each result is drawn at its shortest distance from this item. Arrows run from prerequisites to the items that use them; this item has a heavier outline. External source nodes link directly to the cited text. The view shows up to two dependency steps; open a prerequisite to continue.