OctoInformBeta

OpenAI preprint, September 2026 · §5. Matching the leading transports

Lemma 5.2: Coercivity for smooth coefficients and jointly smooth solvability from a dense parameter set

Compact support controls planar transport solutions.

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

Matching the leading transports in the elasticity inverse problem requires planar solutions with prescribed compact support as parameters vary. Fix a finite component index set II, and write z=x+iyz=x+iy on ℂ\mathbb C.

The test space 𝒟\mathcal D consists of smooth complex-valued functions on ℂ\mathbb C with compact support.✦ The scalar and vector Hilbert spaces in (1) use planar Lebesgue measure dAdA: 𝒟=Cc∞(ℂ;ℂ),L=L2(ℂ;ℂ),H=L2(ℂ;ℂI),∥w∥H2=∑i∈I∥wi∥L2.\mathcal D=C_c^\infty(\mathbb C;\mathbb C),\qquad L=L^2(\mathbb C;\mathbb C),\qquad H=L^2(\mathbb C;\mathbb C^I),\qquad \|w\|_H^2=\sum_{i\in I}\|w_i\|_L^2.(1) Let 𝒮(ℂ)\mathcal S(\mathbb C) denote the Schwartz space on ℂ\mathbb C; the natural inclusions j:𝒟→𝒮(ℂ)j:\mathcal D\to\mathcal S(\mathbb C), ι:𝒟→L\iota:\mathcal D\to L and ιI:𝒟I→H\iota_I:\mathcal D^I\to H are given by (2), where brackets denote an L2L^2 equivalence class: jφ=φ,ιφ=[φ],ιIu=(ιui)i∈I.j\varphi=\varphi,\qquad \iota\varphi=[\varphi],\qquad \iota_Iu=(\iota u_i)_{i\in I}.(2)

For v∈ℂv\in\mathbb C and smooth b:ℂ→ℂb:\mathbb C\to\mathbb C, the real directional derivative ∂v\partial_v, multiplication operator MbM_b and planar Cauchy-Riemann operator BB are defined by (3); the same formula defines BB on Schwartz functions: ∂vφ(z)=Dℝφ(z)[v],Mbφ(z)=b(z)φ(z),Bφ=∂1φ+i∂iφ=2∂‾φ.\partial_v\varphi(z)=D_{\mathbb R}\varphi(z)[v],\qquad M_b\varphi(z)=b(z)\varphi(z),\qquad B\varphi=\partial_1\varphi+i\partial_i\varphi=2\bar\partial\varphi.(3) For a smooth matrix a=(aij)i,j∈Ia=(a_{ij})_{i,j\in I} on ℂ\mathbb C, the transport operator Ta:𝒟I→𝒟IT_a:\mathcal D^I\to\mathcal D^I is the matrix perturbation in (4): (Tau)i=Bui+∑j∈IMaijuj.(T_au)_i=Bu_i+\sum_{j\in I}M_{a_{ij}}u_j.(4)✦

Let 𝒮′(ℂ)\mathcal S'(\mathbb C) denote the space of tempered distributions on ℂ\mathbb C; the component injection Ji:H→𝒮′(ℂ)J_i:H\to\mathcal S'(\mathbb C) associates to w∈Hw\in H the regular tempered distribution defined by wiw_i.✦ A temperately growing coefficient b:ℂ→ℂb:\mathbb C\to\mathbb C is a smooth function whose real derivatives each have at most polynomial growth. For a matrix a=(aij)i,j∈Ia=(a_{ij})_{i,j\in I} of temperately growing coefficients on ℂ\mathbb C, the weak transport operator PaP_a on HH is the vector of tempered distributions given by (5), with derivatives taken distributionally: (Paw)i=B(Jiw)+∑j∈IaijJjw.(P_aw)_i=B(J_iw)+\sum_{j\in I}a_{ij}J_jw.(5)✦ Two tempered distributions FF and GG agree locally on Ω⊂ℂ\Omega\subset\mathbb C when their values agree on every test function supported in Ω\Omega, as in (6): F|Ω=G|Ω⇔F(φ)=G(φ)for every φ∈𝒟 with supp⁡φ⊂Ω.F|_\Omega=G|_\Omega \quad\Longleftrightarrow\quad F(\varphi)=G(\varphi) \quad\text{for every }\varphi\in\mathcal D \text{ with }\operatorname{supp}\varphi\subset\Omega.(6)✦ For a subset SS of ℂ\mathbb C, a vector w∈Hw\in H is weakly supported in SS when JiwJ_iw agrees locally with zero on ScS^c for each i∈Ii\in I.

For a planar set S⊂ℂS\subset\mathbb C and a smooth matrix aa, the supported test space 𝒟S\mathcal D_S, its test transport graph ΓS,a0\Gamma^0_{S,a} and the closed transport graph ΓS,a\Gamma_{S,a} are the complex linear subspaces defined in (7). The closure is taken in H⊕2HH\oplus_2H, whose norm is the square root of the sum of the squared component norms; π2\pi_2 is the second projection restricted to ΓS,a\Gamma_{S,a}: 𝒟S={u∈𝒟I:supp⁡ui⊂S for all i∈I},ΓS,a0={(ιIu,ιITau):u∈𝒟S},ΓS,a=ΓS,a0¯H⊕2H,π2(w,g)=g.\begin{aligned} \mathcal D_S &=\{u\in\mathcal D^I:\operatorname{supp}u_i\subset S\text{ for all }i\in I\},\\ \Gamma^0_{S,a} &=\{(\iota_Iu,\iota_IT_au):u\in\mathcal D_S\},\\ \Gamma_{S,a}&=\overline{\Gamma^0_{S,a}}^{\,H\oplus_2H}, &\pi_2(w,g)&=g. \end{aligned}(7) For a planar set S⊂ℂS\subset\mathbb C and a smooth matrix aa, the test transport graph embedding is the complex linear map from 𝒟S\mathcal D_S to H⊕2HH\oplus_2H sending uu to (ιIu,ιITau)(\iota_Iu,\iota_IT_au); its range is ΓS,a0\Gamma^0_{S,a}.

Let PP be the real parameter space, and let aa and ff be respectively a matrix-valued function and a ℂI\mathbb C^I-valued function on P×ℂP\times\mathbb C; for p∈Pp\in P, write ap(z)=a(p,z)a_p(z)=a(p,z) and fp(z)=f(p,z)f_p(z)=f(p,z).

Test functions on the right side of a distributional equation are identified with their regular distributions.

The Claim

Theorem 1. For a continuous matrix aa and u∈𝒟Iu\in\mathcal D^I satisfying Bui=∑jaijujBu_i=\sum_j a_{ij}u_j for every component, compact support uniqueness is (8): u=0.u=0.(8)✦ For a matrix aa of temperately growing coefficients, compact SS, and w∈Hw\in H weakly supported in SS with Paw=0P_aw=0, weak compact support uniqueness is (9): w=0.w=0.(9)✦ For smooth aa and compact SS, supported transport coercivity (10) holds with some C>0C>0: ∥ιIu∥H≤C∥ιITau∥H(u∈𝒟S).\|\iota_Iu\|_H\le C\|\iota_IT_au\|_H \qquad(u\in\mathcal D_S).(10)✦ Under these hypotheses, closed graph coercivity (11) holds with some CG>0C_G>0: ∥q∥H⊕2H≤CG∥π2q∥H(q∈ΓS,a).\|q\|_{H\oplus_2H}\le C_G\|\pi_2q\|_H \qquad(q\in\Gamma_{S,a}).(11)✦ Let S,K⊂ℂS,K\subset\mathbb C be compact, and let aa and ff be jointly smooth in (p,z)(p,z) and vanish for z∉Kz\notin K. Let D⊂PD\subset P be dense. Suppose that for each p∈Dp\in D there is wp∈Hw_p\in H weakly supported in SS satisfying Papwp=fpP_{a_p}w_p=f_p. Supported parameter inversion (12) gives a jointly smooth function v:P×ℂ→ℂIv:P\times\mathbb C\to\mathbb C^I such that supp⁡vi(p,⋅)⊂S,Bvi(p,z)+∑j∈Iaij(p,z)vj(p,z)=fi(p,z)(p∈P,i∈I,z∈ℂ).\operatorname{supp}v_i(p,\cdot)\subset S,\qquad Bv_i(p,z)+\sum_{j\in I}a_{ij}(p,z)v_j(p,z)=f_i(p,z) \quad(p\in P,\ i\in I,\ z\in\mathbb C).(12)✦

In the paper, the result reads: for an open U⊂ℝdU\subset\mathbb R^d, n≥1n\ge1, a disk radius rr, a parameter ρ∈U\rho\in U and Pρ=γ(ρ)(∂y1+i∂y2)+A(y,ρ)P_\rho=\gamma(\rho)(\partial_{y_1}+i\partial_{y_2})+A(y,\rho), with smooth nonvanishing γ\gamma and smooth complex n×nn\times n matrix AA on a neighborhood of D2r2¯×U\overline{D_{2r}^2}\times U, where DR2={y∈ℝ2:|y|<R}D_R^2=\{y\in\mathbb R^2:|y|<R\}, every distributional homogeneous solution compactly supported in D2r2D_{2r}^2 vanishes; for each integer j≥0j\ge0, every W∈Hj+1(D2r2;ℂn)W\in H^{j+1}(D_{2r}^2;\mathbb C^n) supported in Dr2¯\overline{D_r^2} satisfies ∥W∥Hj+1≤Cj∥PρW∥Hj\|W\|_{H^{j+1}}\le C_j\|P_\rho W\|_{H^j}, locally uniformly in ρ\rho; and, for smooth ff, existence of supported L2L^2 solutions on a dense parameter set implies existence for every parameter, uniqueness and joint smoothness, with all conclusions applying entrywise to matrix unknowns.

What Is Checked

The claims are stated over finite component L2L^2 spaces, compactly supported test functions and their transport graph closures, with joint smoothness and a dense parameter set in (12).

Real directional differentiation and multiplication by a smooth coefficient in (3) preserve 𝒟\mathcal D. A parameter-dependent function that vanishes off a fixed compact set has compactly supported slices.✦