OctoInformBeta

OpenAI preprint, September 2026 · §3. An augmented system and analytic estimates

Lemma 3.4: Sixth-power phase bounds and smooth corrections for the conjugated elasticity equation

Interior control of augmented elastic displacements.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

Conjugated elasticity equations in complex geometrical optics couple a displacement to its shifted divergence. The interior field considered here has four components, with the divergence in the scalar position.

The physical space is X=ℝ3X=\mathbb R^3, with coordinate basis e1,e2,e3e_1,e_2,e_3. The Schwartz space 𝒮(X)\mathcal S(X) consists of smooth complex-valued functions whose derivatives decay faster than every inverse polynomial; its continuous complex-linear dual 𝒮′(X)\mathcal S'(X) is the space of tempered distributions. Distributional and Carleman test functions belong to 𝒮(X)\mathcal S(X). Two tempered distributions agree in a set Ω⊂X\Omega\subset X if they agree on every Schwartz test function with compact support contained in Ω\Omega.✦

The scalar space H=L2(X;ℂ)H=L^2(X;\mathbb C) uses Lebesgue measure.✦ The component spaces HN=HNH_N=H^N, for N=3,4N=3,4, and the Hilbert displacement space ℋ=L2(X;ℂ3)\mathscr H=L^2(X;\mathbb C^3) have the norms in (1): ∥u∥HN=max1≤i≤N∥ui∥L2(X),∥w∥ℋ=(∑i=13∥wi∥L2(X)2)1/2.\|u\|_{H_N}=\max_{1\le i\le N}\|u_i\|_{L^2(X)}, \qquad \|w\|_{\mathscr H}=\left(\sum_{i=1}^3\|w_i\|_{L^2(X)}^2\right)^{1/2}.(1) The scalar Schwartz embedding is the continuous complex-linear inclusion of 𝒮(X)\mathcal S(X) into HH; its componentwise version I3:𝒮(X)3→ℋI_3:\mathcal S(X)^3\to\mathscr H is the vector Schwartz embedding.

The smooth coefficient algebra C∞(X;ℂ)C^\infty(X;\mathbb C) consists of smooth complex-valued functions.✦ The temperate subalgebra 𝒜\mathcal A of C∞(X;ℂ)C^\infty(X;\mathbb C) consists of the coefficients whose derivatives of every order have polynomial growth.✦ A smooth coefficient aa is exterior constant if a−ca-c has compact support for some c∈ℂc\in\mathbb C.✦ In the physical Lamé regime the real coefficients satisfy μ>0\mu>0 and 3λ+2μ>03\lambda+2\mu>0; write m=μm=\mu, k=λ+μk=\lambda+\mu and ℓ=λ+2μ\ell=\lambda+2\mu.

For real vectors α,β∈X\alpha,\beta\in X and h>0h>0, the complex phase vector θ\theta and the exponential function EτE_\tau are given by (2): θ=α+iβ,τ=h−1,z=α+iβh=τθ,Eτ(x)=exp⁡(τθ⋅x).\theta=\alpha+\mathrm i\beta, \qquad \tau=h^{-1}, \qquad z=\frac{\alpha+\mathrm i\beta}{h}=\tau\theta, \qquad E_\tau(x)=\exp(\tau\theta\cdot x).(2)✦ The dot product in the exponential is complex bilinear.

The Distributional System

The directional derivative ∂v\partial_v differentiates Schwartz functions along the real direction vv; ∂i\partial_i denotes differentiation along eie_i, and also its distributional counterpart. For an arbitrary constant z∈ℂ3z\in\mathbb C^3, the shifted derivatives Tz,iT_{z,i} and shifted divergence div⁡z\operatorname{div}_z are the operators in (3): Tz,i=∂i+ziId⁡,div⁡zu=∑i=13Tz,iui.T_{z,i}=\partial_i+z_i\operatorname{Id}, \qquad \operatorname{div}_z u=\sum_{i=1}^3 T_{z,i}u_i.(3)

Multiplication by a temperate coefficient aa on Schwartz space is the continuous complex-linear map Maφ=aφM_a\varphi=a\varphi.✦ For a,b∈𝒜a,b\in\mathcal A, the conjugated elasticity operator Lz(a,b)L_z(a,b) is the shifted stress divergence (4), acting on Schwartz fields or on tempered distributions: (Lz(a,b)u)i=Tz,i(adiv⁡zu)+∑j=13Tz,j(b(Tz,jui+Tz,iuj)),i=1,2,3.\bigl(L_z(a,b)u\bigr)_i =T_{z,i}\bigl(a\operatorname{div}_z u\bigr) +\sum_{j=1}^3T_{z,j}\bigl(b(T_{z,j}u_i+T_{z,i}u_j)\bigr), \qquad i=1,2,3.(4) Write L=L0(λ,μ)L=L_0(\lambda,\mu) for the elasticity operator with physical coefficients. The augmented displacement 𝒰zu∈𝒮′(X)4\mathcal U_z u\in\mathcal S'(X)^4 is the field in (5): 𝒰zu=(div⁡zu,u1,u2,u3).\mathcal U_z u=(\operatorname{div}_z u,u_1,u_2,u_3).(5)

For s∈ℝs\in\mathbb R, the Sobolev injection ιs:H→𝒮′(X)\iota_s:H\to\mathcal S'(X) is the continuous complex-linear map in (6), obtained by viewing an L2L^2 function as a tempered distribution and applying the Bessel potential with Fourier multiplier (1+|2πξ|2)−s/2(1+|2\pi\xi|^2)^{-s/2}.✦ Here ℬt\mathcal B_t denotes the Bessel potential operator with parameter tt. ιsq=ℬ−sq,ιs(N)q=(ιsq1,…,ιsqN),s∈ℝ.\iota_s q=\mathcal B_{-s}q, \qquad \iota_s^{(N)}q=(\iota_s q_1,\ldots,\iota_s q_N), \qquad s\in\mathbb R.(6)

A nested family of five cutoffs in Ω\Omega consists of χ0,…,χ4∈Cc∞(X;ℂ)\chi_0,\ldots,\chi_4\in C_c^\infty(X;\mathbb C) satisfying (7): supp⁡χi⊂Ω(0≤i≤4),χi+1=1on supp⁡χi(0≤i≤3).\operatorname{supp}\chi_i\subset\Omega\quad(0\le i\le4), \qquad \chi_{i+1}=1\ \text{on }\operatorname{supp}\chi_i\quad(0\le i\le3).(7)

The Claim

Theorem 1. Let Ω⊂X\Omega\subset X and let χ0,…,χ4\chi_0,\ldots,\chi_4 be a nested family of cutoffs in Ω\Omega. Let a,b,n,o∈𝒜a,b,n,o\in\mathcal A satisfy nb=1nb=1 and o(a+2b)=1o(a+2b)=1. For z∈ℂ3z\in\mathbb C^3 and r,F∈H3r,F\in H_3 satisfying Lz(a,b)r=FL_z(a,b)r=F distributionally in Ω\Omega, there is q∈H4q\in H_4 satisfying the augmented H1H^1 estimate (8), with a constant C>0C>0 independent of zz, rr and FF: ι1(4)q=χ0𝒰zr,∥q∥H4≤C(1+∑i=13|zi|)6(∥r∥H3+∥F∥H3).\iota_1^{(4)}q=\chi_0\mathcal U_z r, \qquad \|q\|_{H_4}\le C\left(1+\sum_{i=1}^3|z_i|\right)^6 \bigl(\|r\|_{H_3}+\|F\|_{H_3}\bigr).(8)✦

For the classical correction, let R≥1R\ge1, let Ω⊂B¯(0,R)\Omega\subset\overline B(0,R), and let Ω′⊂Ω\Omega'\subset\Omega be open. Let α,β∈X\alpha,\beta\in X with α≠0\alpha\ne0, and let λ,μ\lambda,\mu be real smooth exterior-constant coefficients with μ(x)≠0\mu(x)\ne0 and ℓ(x)≠0\ell(x)\ne0 for every x∈Xx\in X. Let χ0,…,χ4\chi_0,\ldots,\chi_4 be nested cutoffs in Ω\Omega with χ0=1\chi_0=1 on Ω′\Omega'. For f∈𝒮(X)3f\in\mathcal S(X)^3 and sufficiently small h>0h>0, there exist w∈ℋw\in\mathscr H, q∈H4q\in H_4 and v∈𝒮(X)3v\in\mathcal S(X)^3 with the properties (9), where zz is as in (2) and the constants C,h0>0C,h_0>0 are independent of hh and ff, the admissible range being 0<h≤h00<h\le h_0: w=valmost everywhere on Ω′,Lz(λ,μ)v=fpointwise on Ω′,ι1(4)q=χ0𝒰zw,∥w∥ℋ≤C∥I3f∥ℋ,∥q∥H4≤Ch−6∥I3f∥ℋ.\begin{gathered} w=v\quad\text{almost everywhere on }\Omega', \qquad L_z(\lambda,\mu)v=f\quad\text{pointwise on }\Omega',\\ \iota_1^{(4)}q=\chi_0\mathcal U_z w, \qquad \|w\|_{\mathscr H}\le C\|I_3f\|_{\mathscr H}, \qquad \|q\|_{H_4}\le Ch^{-6}\|I_3f\|_{\mathscr H}. \end{gathered}(9)✦

In the paper, the result reads: for B′⋐BB'\Subset B, there are CB′,hB′>0C_{B'},h_{B'}>0 such that every u,f∈L2(B;ℂ3)u,f\in L^2(B;\mathbb C^3) solving Eτ−1L(Eτu)=fE_\tau^{-1}L(E_\tau u)=f distributionally, with 0<h=τ−1<hB′0<h=\tau^{-1}<h_{B'}, satisfies (10), ∑j=02hj∥∇ju∥L2(B′)≤CB′(∥u∥L2(B)+h2∥f∥L2(B)),\sum_{j=0}^2h^j\|\nabla^j u\|_{L^2(B')} \le C_{B'}\bigl(\|u\|_{L^2(B)}+h^2\|f\|_{L^2(B)}\bigr),(10) and smooth ff gives smooth uu in BB for each fixed τ\tau.

What Is Checked

The augmented estimate is stated over temperate coefficients with temperate inverses of bb and a+2ba+2b, componentwise L2L^2 fields and the Sobolev injection into tempered distributions; the classical correction uses the Hilbert displacement space and real exterior-constant coefficients with the stated nonvanishing conditions. Every exterior-constant smooth coefficient has temperate growth.✦