OctoInformBeta

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

Lemma 5.1: A pointwise amplitude change and its real transport coefficients

Normalizing the leading elasticity transports.

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

The leading displacement amplitude in the augmented isotropic elasticity system is coupled to a scalar amplitude through derivatives of the Lamé coefficients. For real coefficients, write m=μm=\mu, k=λ+μk=\lambda+\mu and ℓ=λ+2μ\ell=\lambda+2\mu, with μ>0\mu>0 and 3λ+2μ>03\lambda+2\mu>0. The transport equations differentiate the amplitudes along a complex direction θ∈ℂ3\theta\in\mathbb C^3.

The coefficient algebra C∞(X;ℂ)C^\infty(X;\mathbb C) consists of smooth complex-valued functions on X=ℝ3X=\mathbb R^3, with pointwise addition and multiplication.✦ Real smooth functions are regarded as complex-valued by the natural inclusion.✦ A nowhere-zero coefficient ff has the smooth reciprocal f−1(x)=1/f(x)f^{-1}(x)=1/f(x).✦

For a smooth complex-valued function ff on XX, a real vector v∈ℝ3v\in\mathbb R^3 and the standard coordinate vectors eie_i, the directional derivative DvD_v and the coordinate derivatives ∂i\partial_i are given in (1); the differential is the real differential, and the product of θ\theta with a vector z∈ℂ3z\in\mathbb C^3 is complex bilinear, without conjugation. Dvf(x)=dfx(v),∂if=Deif,Dθf=∑i=13θi∂if,θ⋅z=∑i=13θizi.D_v f(x)=\mathrm df_x(v),\qquad \partial_i f=D_{e_i}f,\qquad D_\theta f=\sum_{i=1}^3\theta_i\partial_i f,\qquad \theta\cdot z=\sum_{i=1}^3\theta_i z_i.(1) For a nonzero null direction, the constrained transport fiber is the complex vector space EθE_\theta in (2). θ≠0,θ⋅θ=0,Hθ={a∈ℂ3:θ⋅a=0},Eθ=Hθ⊕ℂ.\theta\ne0,\qquad \theta\cdot\theta=0,\qquad H_\theta=\{a\in\mathbb C^3:\theta\cdot a=0\},\qquad E_\theta=H_\theta\oplus\mathbb C.(2)

The Amplitude Change

Let k,rk,r be complex smooth coefficients, and put m=r2m=r^2, λ=k−r2\lambda=k-r^2 and ℓ=k+r2\ell=k+r^2. The gradient ∇f\nabla f of a smooth complex-valued function ff on XX is the vector of its three coordinate derivatives. The leading vector transport operator Rθ[k,r]R_\theta[k,r] and scalar transport operator Sθ[k,r]S_\theta[k,r] act on a smooth vector amplitude aa and scalar amplitude bb by (3). Rθ[k,r](a,b)=2mDθa+(Dθm)a+θ(kb+∇m⋅a),Sθ[k,r](a,b)=2ℓDθb+2(Dθk)b+4∇m⋅Dθa+2(Dθ∇m)⋅a.\begin{aligned} R_\theta[k,r](a,b) &=2mD_\theta a+(D_\theta m)a +\theta\bigl(kb+\nabla m\cdot a\bigr),\\ S_\theta[k,r](a,b) &=2\ell D_\theta b+2(D_\theta k)b +4\nabla m\cdot D_\theta a +2(D_\theta\nabla m)\cdot a. \end{aligned}(3)

The logarithmic gradient Γr\Gamma_r and its directional contraction γθ[r]\gamma_\theta[r] for nowhere-zero rr are the smooth fields in (4). Γr=2r−1∇r,γθ[r]=θ⋅Γr.\Gamma_r=2r^{-1}\nabla r,\qquad \gamma_\theta[r]=\theta\cdot\Gamma_r.(4) For a nowhere-zero smooth complex coefficient rr and a smooth complex coefficient kk on XX, the vector and scalar amplitude transformations 𝒫r\mathcal P_r and 𝒮k,r\mathcal S_{k,r} send smooth vector and scalar amplitudes a,ba,b to the amplitudes p,sp,s given by (5). p=𝒫r(a)=ra,s=𝒮k,r(a,b)=−12(krb+Γr⋅p).p=\mathcal P_r(a)=ra,\qquad s=\mathcal S_{k,r}(a,b) =-\frac12\left(\frac{k}{r}b+\Gamma_r\cdot p\right).(5) For nowhere-zero smooth complex coefficients k,rk,r on XX and smooth vector and scalar amplitudes p,sp,s, the recovered vector and scalar amplitudes 𝒜r(p)\mathcal A_r(p) and ℬk,r(p,s)\mathcal B_{k,r}(p,s) are the smooth functions given by (6). 𝒜r(p)=r−1p,ℬk,r(p,s)=−rk(2s+Γr⋅p).\mathcal A_r(p)=r^{-1}p,\qquad \mathcal B_{k,r}(p,s) =-\frac{r}{k}\bigl(2s+\Gamma_r\cdot p\bigr).(6)

The smooth transport coefficient 𝒯θ[k,r]\mathcal T_\theta[k,r] and vector 𝒬θ[k,r]\mathcal Q_\theta[k,r] for nowhere-zero k,r,ℓk,r,\ell are defined by (7). 𝒯θ[k,r]=Dθkk−Dθℓℓ,𝒬θ[k,r]=12[(r2ℓDθkk−12γθ[r])Γr−r2ℓDθΓr].\begin{aligned} \mathcal T_\theta[k,r] &=\frac{D_\theta k}{k}-\frac{D_\theta\ell}{\ell},\\ \mathcal Q_\theta[k,r] &=\frac12\left[ \left(\frac{r^2}{\ell}\frac{D_\theta k}{k} -\frac12\gamma_\theta[r]\right)\Gamma_r -\frac{r^2}{\ell}D_\theta\Gamma_r\right]. \end{aligned}(7)

The gradient ∇f\nabla f of a real smooth function ff on XX is a real vector field and its Hessian ∇2f\nabla^2 f is a real matrix field, with components given by (8). (∇f)i=∂if,(∇2f)ij=∂j∂if.(\nabla f)_i=\partial_i f,\qquad (\nabla^2 f)_{ij}=\partial_j\partial_i f.(8) For real smooth positive m,km,k, let g=∇log⁡mg=\nabla\log m and ℓ=k+m\ell=k+m. The positive square root r=mr=\sqrt m is regarded as a complex smooth coefficient.✦ The real transport matrix M[m,k](x)∈ℝ3×3M[m,k](x)\in\mathbb R^{3\times3} and the recovery transport coefficients Tθ[m,k]T_\theta[m,k], Qθ[m,k]Q_\theta[m,k] are defined by (9). M[m,k]ij=12[(mℓ∂jlogk−12gj)gi−mℓ(∇2logm)ij],Tθ[m,k]=Dθlog⁡kℓ,Qθ[m,k]=M[m,k]θ.\begin{aligned} M[m,k]_{ij} &=\frac12\left[ \left(\frac{m}{\ell}\partial_j\log k-\frac12g_j\right)g_i -\frac{m}{\ell}(\nabla^2\log m)_{ij}\right],\\ T_\theta[m,k]&=D_\theta\log\frac{k}{\ell},\qquad Q_\theta[m,k]=M[m,k]\theta. \end{aligned}(9)

The Claim

Theorem 1. Let k,r∈C∞(X;ℂ)k,r\in C^\infty(X;\mathbb C) be such that rr, kk and ℓ=k+r2\ell=k+r^2 are nowhere zero. Let a∈C∞(X;ℂ3)a\in C^\infty(X;\mathbb C^3) and b∈C∞(X;ℂ)b\in C^\infty(X;\mathbb C), and set p=𝒫r(a)p=\mathcal P_r(a) and s=𝒮k,r(a,b)s=\mathcal S_{k,r}(a,b). For every θ∈ℂ3\theta\in\mathbb C^3 and x∈Xx\in X, the leading equations have the pointwise equivalence (10). Rθ[k,r](a,b)(x)=0,Sθ[k,r](a,b)(x)=0⇔{Dθp(x)=θs(x),Dθs(x)=𝒯θ[k,r](x)s(x)+𝒬θ[k,r](x)⋅p(x).\begin{aligned} &R_\theta[k,r](a,b)(x)=0,\quad S_\theta[k,r](a,b)(x)=0\\ &\qquad\Longleftrightarrow\qquad \begin{cases} D_\theta p(x)=\theta s(x),\\ D_\theta s(x)=\mathcal T_\theta[k,r](x)s(x) +\mathcal Q_\theta[k,r](x)\cdot p(x). \end{cases} \end{aligned}(10)✦ For arbitrary smooth vector and scalar amplitudes π\pi and σ\sigma, the recovery maps satisfy (11) and (12). 𝒫r(𝒜r(π))=π.\mathcal P_r\bigl(\mathcal A_r(\pi)\bigr)=\pi.(11)✦ 𝒮k,r(𝒜r(π),ℬk,r(π,σ))=σ.\mathcal S_{k,r}\bigl(\mathcal A_r(\pi), \mathcal B_{k,r}(\pi,\sigma)\bigr)=\sigma.(12)✦ For real smooth positive coefficients m,km,k and r=mr=\sqrt m, the coefficient identities are (13) and (14), at every x∈Xx\in X and for each component ii. 𝒯θ[k,m](x)=Tθ[m,k](x).\mathcal T_\theta[k,\sqrt m](x)=T_\theta[m,k](x).(13)✦ 𝒬θ,i[k,m](x)=Qθ,i[m,k](x).\mathcal Q_{\theta,i}[k,\sqrt m](x)=Q_{\theta,i}[m,k](x).(14)✦

In the paper, the result reads: for one common extended coefficient pair, with g=∇log⁡mg=\nabla\log m, the change p=map=\sqrt m\,a and s=−kb/(2m)−(g⋅p)/2s=-kb/(2\sqrt m)-(g\cdot p)/2 is a smooth invertible map on EθE_\theta which transforms the constrained leading equations into (15), and both the derivative and the transport matrix depend linearly and holomorphically on θ\theta: Dθ(ps)=(0θQθ⋅Tθ)(ps),Tθ=Dθlog⁡(k/ℓ),Qθ=12[(mℓDθlogk−12θ⋅g)g−mℓDθg].\begin{aligned} D_\theta\binom{p}{s} &=\begin{pmatrix}0&\theta\\Q_\theta\cdot&T_\theta\end{pmatrix} \binom{p}{s},\\ T_\theta&=D_\theta\log(k/\ell),\qquad Q_\theta=\frac12\left[ \left(\frac m\ell D_\theta\log k-\frac12\theta\cdot g\right)g -\frac m\ell D_\theta g\right]. \end{aligned}(15)

What Is Checked

The pointwise equivalence is stated over smooth complex amplitudes and coefficients on ℝ3\mathbb R^3, with arbitrary complex θ\theta and nowhere-zero r,k,k+r2r,k,k+r^2; the real specialization uses positive smooth m,km,k. For these real coefficients, the positive square root and k+(m)2k+(\sqrt m)^2 are nowhere zero.