OctoInformBeta

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

Lemma 3.1: Exact divergence and forcing identities accompany local formulas for the lower-order coefficients

Elasticity admits a normalized augmented system.

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

In the isotropic inverse boundary problem, adjoining an independent scalar to the elastic displacement makes the divergence constraint part of the differential system. The questions are its compatibility with the physical equation and the principal part of the coupled operator.

The spatial domain is X=ℝ3X=\mathbb R^3, and the smooth coefficient algebra is C∞(X;ℂ)C^\infty(X;\mathbb C) with pointwise operations. For real Lamé coefficients λ,μ\lambda,\mu, use μ>0\mu>0 and 3λ+2μ>03\lambda+2\mu>0, and write m=μm=\mu, k=λ+mk=\lambda+m and ℓ=λ+2m\ell=\lambda+2m. The differential constructions also take complex smooth λ,m\lambda,m.

Let eie_i be the iith coordinate basis vector, with 1≤i≤31\leq i\leq3. The coordinate derivative ∂i\partial_i is differentiation in direction eie_i; ∂v=v⋅∇\partial_v=v\cdot\nabla denotes differentiation in a real direction vv. For a smooth vector field uu and a smooth scalar function ff on XX, the divergence and Laplacian are the operators in (1): div⁡u=∑i=13∂iui,Δf=∑i=13∂i∂if.\operatorname{div}u=\sum_{i=1}^3\partial_i u_i, \qquad \Delta f=\sum_{i=1}^3\partial_i\partial_i f.(1)

Residuals and Normalization

For operators TiT_i on smooth functions, the associated Laplacian ΔT\Delta_T is defined in (2). For a constant vector z∈ℂ3z\in\mathbb C^3, the same display defines the shifted derivatives ∂iz\partial_i^z and shifted divergence div⁡z\operatorname{div}_z; the dot product of complex vectors is bilinear, without conjugation. ΔTf=∑i=13Ti(Tif),∂izf=(∂i+zi)f,div⁡zu=∑i=13∂izui.\Delta_T f=\sum_{i=1}^3T_i(T_i f), \qquad \partial_i^z f=(\partial_i+z_i)f, \qquad \operatorname{div}_z u=\sum_{i=1}^3\partial_i^z u_i.(2)

The shifted shear modulus gradient coupling CT(m,u)C_T(m,u) is the vector field in (3), for a smooth displacement uu; its unshifted version is C=C∂C=C_{\partial}. CT(m,u)i=∑j=13(∂jm)(Tjui+Tiuj).C_T(m,u)_i=\sum_{j=1}^3(\partial_jm)(T_ju_i+T_iu_j).(3)✦

The shifted displacement residual RT(u,d)R_T(u,d) and scalar augmented residual ST(u,d)S_T(u,d) are the vector and scalar fields defined in (4), for a smooth pair (u,d)(u,d); write R=R∂R=R_{\partial} and S=S∂S=S_{\partial}. RT(u,d)i=mΔTui+kTid+CT(m,u)i+(∂iλ)d,ST(u,d)=ℓΔTd+2∑i(∂ik)Tid+2∑i(∂im)ΔTui+2∑i,j(∂i∂jm)Tiuj+(Δλ)d.\begin{aligned} R_T(u,d)_i &=m\Delta_Tu_i+kT_id+C_T(m,u)_i+(\partial_i\lambda)d,\\ S_T(u,d) &=\ell\Delta_Td+2\sum_i(\partial_i k)T_id +2\sum_i(\partial_i m)\Delta_Tu_i\\ &\quad+2\sum_{i,j}(\partial_i\partial_jm)T_iu_j +(\Delta\lambda)d. \end{aligned}(4)

The scalar divergence operator PmP_m on smooth functions is defined by (5). Pmf=mΔf+∑i(∂im)(∂if).P_m f=m\Delta f+\sum_i(\partial_i m)(\partial_i f).(5)✦

For smooth auxiliary coefficients n,on,o, the normalized augmented operator 𝒩T\mathcal N_T is the four-component operator in (6). Its argument is ordered as U=(d,u1,u2,u3)U=(d,u_1,u_2,u_3), with scalar component first. 𝒩TU=(o(ST(u,d)−2n∑i(∂im)RT(u,d)i)nRT(u,d)).\mathcal N_T U= \begin{pmatrix} o\bigl(S_T(u,d)-2n\sum_i(\partial_i m)R_T(u,d)_i\bigr)\\ nR_T(u,d) \end{pmatrix}.(6)✦

The first-order matrices ArA_r and potential matrix VV, with rows and columns indexed by 0,1,2,30,1,2,3, are the smooth matrices specified in (7); δij\delta_{ij} denotes the Kronecker symbol, and i,j,ri,j,r range from 11 to 33. (Ar)00=2o(∂rk−kn∂rm),(Ar)0j=2o(∂r∂jm−2n(∂rm)(∂jm)),(Ar)i0=δrikn,(Ar)ij=δjin∂rm+δrin∂jm,V00=o(Δλ−2n∑i(∂im)(∂iλ)),Vi0=n∂iλ,V0j=Vij=0.\begin{aligned} (A_r)_{00}&=2o\bigl(\partial_r k-kn\partial_r m\bigr),& (A_r)_{0j}&=2o\bigl(\partial_r\partial_jm -2n(\partial_r m)(\partial_jm)\bigr),\\ (A_r)_{i0}&=\delta_{ri}kn,& (A_r)_{ij}&=\delta_{ji}n\partial_r m+\delta_{ri}n\partial_jm,\\ V_{00}&=o\bigl(\Delta\lambda -2n\sum_i(\partial_i m)(\partial_i\lambda)\bigr),& V_{i0}&=n\partial_i\lambda,\qquad V_{0j}=V_{ij}=0. \end{aligned}(7)

The Conjugated Physical System

The test space is the complex Schwartz space 𝒮(X)\mathcal S(X); its directional derivatives and Laplacian are those of (1). For smooth multipliers aa of temperate growth, coefficient multiplication on this space is denoted by Maf=afM_a f=af.✦ The conjugated elasticity operator LzL_z is the operator on 𝒮(X)3\mathcal S(X)^3 defined in (8); write L=L0L=L_0. (Lzu)i=∂iz(λdiv⁡zu)+∑j∂jz(m(∂jzui+∂izuj)).(L_z u)_i =\partial_i^z\bigl(\lambda\operatorname{div}_z u\bigr) +\sum_j\partial_j^z\bigl(m(\partial_j^z u_i+\partial_i^z u_j)\bigr).(8)✦

For h∈ℝ\{0}h\in\mathbb R\setminus\{0\} and α,β∈X\alpha,\beta\in X, the phase vector is z=(α+iβ)/hz=(\alpha+\mathrm i\beta)/h.✦ Put θ=α+iβ\theta=\alpha+\mathrm i\beta. The scaled shifted derivative 𝒟h,i\mathscr D_{h,i} and negative conjugated Laplacian 𝒫h\mathscr P_h on 𝒮(X)\mathcal S(X) are defined in (9). 𝒟h,if=h2∂if+hθif,𝒫hf=−h2Δf−2hθ⋅∇f−(θ⋅θ)f.\begin{aligned} \mathscr D_{h,i}f&=h^2\partial_i f+h\theta_i f,\\ \mathscr P_hf&=-h^2\Delta f-2h\theta\cdot\nabla f -(\theta\cdot\theta)f. \end{aligned}(9)

For scalar derivative coefficients a=(a1,a2,a3)a=(a_1,a_2,a_3) and a potential vv, the conjugated lower-order operator ℒha,v\mathscr L_h^{a,v} is defined in (10); for matrix coefficients ar,va_r,v, the same display defines the conjugated augmented system 𝒬ha,v\mathscr Q_h^{a,v} on 𝒮(X)4\mathcal S(X)^4, with 0≤p,q≤30\leq p,q\leq3. ℒha,vf=∑rMar𝒟h,rf+h2Mvf,(𝒬ha,vU)q=𝒫hUq+∑p=03ℒha⋅qp,vqpUp.\begin{aligned} \mathscr L_h^{a,v}f &=\sum_r M_{a_r}\mathscr D_{h,r}f+h^2M_vf,\\ (\mathscr Q_h^{a,v}U)_q &=\mathscr P_hU_q +\sum_{p=0}^3\mathscr L_h^{a_{\cdot qp},v_{qp}}U_p. \end{aligned}(10)

The physical Carleman derivative matrices ara_r, potential matrix vv and forcing correction cc are defined by (11). ar=−Ar,v=−V,ci=−2n∂im.a_r=-A_r,\qquad v=-V,\qquad c_i=-2n\partial_i m.(11)

The normalized augmented forcing ℱhF\mathscr F_hF associated with a vector forcing F∈𝒮(X)3F\in\mathcal S(X)^3 is the four-component field in (12). ℱhF=(o∑i(𝒟h,iFi+h2ciFi)h2nF).\mathscr F_hF= \begin{pmatrix} o\sum_i\bigl(\mathscr D_{h,i}F_i+h^2c_iF_i\bigr)\\ h^2nF \end{pmatrix}.(12)✦

The Claim

Theorem 1. Let λ,m\lambda,m be smooth complex coefficients on XX. For every smooth vector field uu and scalar field dd, the divergence defect identity is div⁡R(u,d)−S(u,d)=Pm(div⁡u−d).\operatorname{div}R(u,d)-S(u,d) =P_m(\operatorname{div}u-d).(13)✦ Let n,on,o be smooth coefficients satisfying nm=1nm=1 and oℓ=1o\ell=1. For every family TiT_i of operators on smooth functions and smooth four-component field UU, the matrix form is 𝒩TU=ΔTU+∑r=13ArTrU+VU.\mathcal N_TU=\Delta_TU+\sum_{r=1}^3A_rT_rU+VU.(14)✦ Fix h∈ℝ\{0}h\in\mathbb R\setminus\{0\} and α,β∈X\alpha,\beta\in X, and take z=(α+iβ)/hz=(\alpha+\mathrm i\beta)/h. Suppose that λ,m,n,o\lambda,m,n,o, every entry of Ar,VA_r,V and every cic_i have temperate growth. For every u∈𝒮(X)3u\in\mathcal S(X)^3, the physical-to-augmented identity is 𝒬h−A,−V(div⁡zuu)=−ℱh(Lzu).\mathscr Q_h^{-A,-V} \begin{pmatrix}\operatorname{div}_z u\\u\end{pmatrix} =-\mathscr F_h(L_z u).(15)✦ Finally, let B⊂XB\subset X be open, and take two smooth coefficient families (λν,mν,nν,oν)(\lambda_\nu,m_\nu,n_\nu,o_\nu), ν=1,2\nu=1,2, satisfying nνmν=1n_\nu m_\nu=1 and oν(λν+2mν)=1o_\nu(\lambda_\nu+2m_\nu)=1. If λ1=λ2\lambda_1=\lambda_2 and m1=m2m_1=m_2 on BB, their normalized matrices satisfy Ar(1)|B=Ar(2)|B(1≤r≤3),V(1)|B=V(2)|B.A_r^{(1)}\big|_B=A_r^{(2)}\big|_B\quad(1\leq r\leq3), \qquad V^{(1)}\big|_B=V^{(2)}\big|_B.(16)✦

In the paper, the result reads: for every smooth pair (u,d)(u,d), Lu=R(u,div⁡u)Lu=R(u,\operatorname{div}u) and div⁡R(u,d)−S(u,d)=Pm(div⁡u−d)\operatorname{div}R(u,d)-S(u,d)=P_m(\operatorname{div}u-d), hence Lu=0Lu=0 implies R(u,div⁡u)=S(u,div⁡u)=0R(u,\operatorname{div}u)=S(u,\operatorname{div}u)=0; the normalized operator is ΔI4+∑rAr(x)∂r+V(x)\Delta I_4+\sum_rA_r(x)\partial_r+V(x) with smooth coefficients, where I4I_4 denotes the identity matrix on ℂ4\mathbb C^4, and for the two common extensions from a domain Ω\Omega its difference has order at most one with coefficients supported in Ω¯\overline\Omega.

What Is Checked

The identities (13) and (14) are stated over smooth coefficients and fields with reciprocal coefficients for normalization, (15) is stated over Schwartz displacements with the specified temperate growth hypotheses, and (16) concerns two smooth coefficient families agreeing on an open set.

The smooth coefficient algebra contains constants and is closed under addition and multiplication, and coordinate differentiation is complex linear and satisfies the Leibniz rule. On 𝒮(X)\mathcal S(X), directional differentiation, shifted differentiation and the Laplacian are complex linear, while coefficient multiplication is continuous and complex linear.