OctoInformBeta

OpenAI preprint, September 2026 · Boundary kernels and the multiplier identity

Lemma 4.3: scalar bulk boundary energy: Harmonic energy, a resolvent pairing, and a boundary form for conformally transported scalar data

Scalar extension energy from boundary data.

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

Scalar bulk energy pairs two gradients and subtracts the potential pairing of the functions. Boundary values on a planar domain are compared on the unit circle through a conformal parametrization. Work in the Euclidean plane ℝ2\mathbb R^2, identified with ℂ\mathbb C by the orthonormal coordinates (1,i)(1,i). The unit disk is D={x:|x|<1}D=\{x:|x|<1\} and its boundary is S1={s:|s|=1}S^1=\{s:|s|=1\}. Arc length dsds on S1S^1 is the image of angular Lebesgue measure under t↦eitt\mapsto e^{it}, with tt taken modulo 2π2\pi; a real boundary function hh has angular datum t↦h(eit)t\mapsto h(e^{it}), regarded also as a complex function.

For an open set UU, a Sobolev function v∈H1(U)v\in H^1(U) is a square-integrable scalar function with a square-integrable weak gradient gg.✦ The weak gradient graph has the Euclidean product norm of these two L2L^2 components. Its defining identity is (1), where φ∈Cc∞(U)\varphi\in C_c^\infty(U) and e∈ℝ2e\in\mathbb R^2; ∂eφ\partial_e\varphi and φe\varphi e are the corresponding scalar derivative and test vector field. ∫Uv∂eφdx=−∫U(g⋅e)φdx,∇v=g.\int_U v\,\partial_e\varphi\,dx =-\int_U (g\cdot e)\varphi\,dx, \qquad \nabla v=g.(1) The test space Cc∞(U)C_c^\infty(U) consists of smooth real functions with compact closed support contained in UU.✦ The inclusion of Cc∞(U)C_c^\infty(U) in H1(U)H^1(U) uses the classical gradient; the space H01(U)H_0^1(U) is the closure of Cc∞(U)C_c^\infty(U) in H1(U)H^1(U). The value and gradient maps on these Sobolev spaces take their values in scalar and vector L2(U)L^2(U), respectively.

Let UU be bounded and let νU\nu_U be a measure dominated by CUdx|UC_U\,dx|_U, with 0≤CU<∞0\leq C_U<\infty. The potential value map J:H1(U)→L2(νU)J:H^1(U)\to L^2(\nu_U) regards a Sobolev function as its value class for νU\nu_U; write J0J_0 for its restriction to H01(U)H_0^1(U). The Dirichlet form aUa_U and weighted Poisson solver 𝒮U\mathcal S_U are given by (2), with datum f∈L2(νU)f\in L^2(\nu_U). Here 𝒮U\mathcal S_U is the variational solution operator for aUa_U and J0J_0, and TUT_U is its abstract Green operator. aU(v,w)=∫U∇v⋅∇wdx,aU(𝒮Uf,v)=⟨f,J0v⟩L2(νU),v∈H01(U),TU=J0𝒮U.\begin{aligned} a_U(v,w)&=\int_U\nabla v\cdot\nabla w\,dx,\\ a_U(\mathcal S_U f,v)&=\langle f,J_0v\rangle_{L^2(\nu_U)}, \qquad v\in H_0^1(U),\\ T_U&=J_0\mathcal S_U. \end{aligned}(2) A function A∈H1(U)A\in H^1(U) is weakly harmonic when aU(A,v)=0a_U(A,v)=0 for every v∈H01(U)v\in H_0^1(U), with the same gradient pairing defining aU(A,v)a_U(A,v).✦ For a bounded real linear operator SS on a real Banach space, the Neumann series resolvent ℛ(S)\mathscr R(S) is the operator series in (3). The subcritical extension ℰU:H1(U)→H1(U)\mathscr E_U:H^1(U)\to H^1(U) is defined there as well, with the correction included from H01(U)H_0^1(U) into H1(U)H^1(U). ℛ(S)=∑n=0∞Sn,ℰUA=A+𝒮Uℛ(TU)JA.\mathscr R(S)=\sum_{n=0}^{\infty}S^n, \qquad \mathscr E_U A=A+\mathcal S_U\mathscr R(T_U)JA.(3)

Conformal Data and Boundary Kernels

Let Ω\Omega be a nonempty bounded simply connected planar domain with smooth boundary. A planar domain Ω\Omega has smooth boundary if near every boundary point pp there is a smooth real defining function ρ\rho, with ρ(p)=0\rho(p)=0 and dρ(p)≠0d\rho(p)\ne0, for which Ω\Omega is locally {ρ<0}\{\rho<0\}.✦ Its first positive Neumann value μ\mu is the infimum of the mean-zero Rayleigh quotients in (4). μ=inf⁡{∫Ω|∇v|2dx∫Ωv2dx:v∈H1(Ω),∫Ωvdx=0,∫Ωv2dx>0}.\mu=\inf\left\{ \frac{\int_\Omega|\nabla v|^2\,dx}{\int_\Omega v^2\,dx}: v\in H^1(\Omega),\quad \int_\Omega v\,dx=0,\quad \int_\Omega v^2\,dx>0\right\}.(4)✦

A closed disk conformal chart consists of a smooth conformal diffeomorphism Φ:D→Ω\Phi:D\to\Omega, with smooth inverse, and globally Lipschitz extensions F,Ψ:ℝ2→ℝ2F,\Psi:\mathbb R^2\to\mathbb R^2 of Φ,Φ−1\Phi,\Phi^{-1}. These extensions map the respective closures into one another and are mutual inverses there. The differentials of Φ\Phi and Φ−1\Phi^{-1} are complex multiplication by nowhere-zero factors dd and d′d', bounded in modulus by positive constants MM and MinvM_{\mathrm{inv}} on their respective domains. The conformal area measure mΦm_\Phi, the potential measure ν\nu, and its density bound CΦC_\Phi are defined in (5), with r+=max⁡{r,0}r_+=\max\{r,0\}. dmΦ=|d(x)|2dx|D,ν=μ+mΦ,CΦ=(μM2)+.dm_\Phi=|d(x)|^2\,dx|_D, \qquad \nu=\mu_+m_\Phi, \qquad C_\Phi=(\mu M^2)_+.(5)

The disk Dirichlet Green kernel GG, Poisson kernel PP, and boundary interaction N0N_0 are the real kernels in (6), with x,y∈Dx,y\in D and s,t∈S1s,t\in S^1, using the same formulas in complex or planar coordinates. G(x,y)=14πlog⁡(1+(1−|x|2)(1−|y|2)|x−y|2),P(s,x)=1−|x|22π|s−x|2,N0(s,t)=1π|s−t|2.\begin{aligned} G(x,y)&=\frac1{4\pi}\log\left( 1+\frac{(1-|x|^2)(1-|y|^2)}{|x-y|^2}\right),\\ P(s,x)&=\frac{1-|x|^2}{2\pi|s-x|^2}, &N_0(s,t)&=\frac1{\pi|s-t|^2}. \end{aligned}(6) The Green integral operator 𝒢\mathcal G acts on L2(D,dx)L^2(D,dx) by integration against GG.✦ The weighted inclusion I:L2(D,dx)→L2(ν)I:L^2(D,dx)\to L^2(\nu) is the continuous change-of-measure map.✦ Its Hilbert adjoint gives the weighted Green operator TT in (7). (𝒢f)(x)=∫DG(x,y)f(y)dy,T=I𝒢I*.(\mathcal G f)(x)=\int_DG(x,y)f(y)\,dy, \qquad T=I\mathcal G I^*.(7)

For Lipschitz real data hh on S1S^1, the real Poisson extension AhA_h is the real part of the complex Poisson extension of its angular datum.✦ Its Sobolev class and its subcritical disk extension UhU_h are denoted as in (8). Ah(x)=∫S1P(s,x)h(s)ds,Ah∈H1(D),Uh=ℰDAh,A_h(x)=\int_{S^1}P(s,x)h(s)\,ds, \qquad A_h\in H^1(D), \qquad U_h=\mathscr E_D A_h,(8) where ℰD\mathscr E_D uses the potential measure ν\nu. The smoothed Poisson kernel P1P_1 and its L2(ν)L^2(\nu) row psp_s are defined by (9). P1(s,x)=∫DG(x,y)P(s,y)dν(y),ps=[P1(s,⋅)]L2(ν).P_1(s,x)=\int_DG(x,y)P(s,y)\,d\nu(y), \qquad p_s=[P_1(s,\cdot)]_{L^2(\nu)}.(9) The Poisson overlap R0R_0, smoothed Poisson overlap R1R_1, and resolvent boundary interaction R2R_2 give the remainder kernel RR in (10). R0(s,t)=∫DP(s,x)P(t,x)dν(x),R1(s,t)=∫DP(s,x)P1(t,x)dν(x),R2(s,t)=⟨ps,ℛ(T)pt⟩L2(ν),R(s,t)=R0(s,t)+R1(s,t)+R2(s,t).\begin{aligned} R_0(s,t)&=\int_DP(s,x)P(t,x)\,d\nu(x),\\ R_1(s,t)&=\int_DP(s,x)P_1(t,x)\,d\nu(x),\\ R_2(s,t)&=\langle p_s,\mathscr R(T)p_t\rangle_{L^2(\nu)},\\ R(s,t)&=R_0(s,t)+R_1(s,t)+R_2(s,t). \end{aligned}(10)

For a real function ww on the plane, the pulled datum is w∘Fw\circ F. If ww is continuous on Ω¯\overline\Omega, its pulled boundary trace hw(s)=w(F(s))h_w(s)=w(F(s)) is a continuous real function on S1S^1. For the scalar solutions, take ww smooth on Ω¯\overline\Omega. Sobolev transport 𝒯:H01(D)→H01(Ω)\mathcal T:H_0^1(D)\to H_0^1(\Omega) is the continuous extension of test function transport: a test function ff is sent to f∘Φ−1f\circ\Phi^{-1} on Ω\Omega, extended by zero outside Ω\Omega.✦ For μ≥0\mu\geq0, the zero-trace datum correction and transported physical solution are specified by (11); when μ>0\mu>0, the transported solution uw∈H1(Ω)u_w\in H^1(\Omega) is the scalar Dirichlet solution. δw=Uhw−(w∘F)∈H01(D),uw=w+𝒯δw∈H1(Ω).\delta_w=U_{h_w}-(w\circ F)\in H_0^1(D), \qquad u_w=w+\mathcal T\delta_w\in H^1(\Omega).(11)

The Claim

Theorem 1 (Scalar bulk and boundary energy). Let UU be bounded and open, let νU≤CUdx|U\nu_U\leq C_U\,dx|_U with 0≤CU<∞0\leq C_U<\infty, and suppose ∥TU∥<1\|T_U\|<1. For weakly harmonic A,B∈H1(U)A,B\in H^1(U), the extensions satisfy (12). ∫U∇(ℰUA)⋅∇(ℰUB)dx−⟨JℰUA,JℰUB⟩L2(νU)=∫U∇A⋅∇Bdx−⟨JA,ℛ(TU)JB⟩L2(νU).\begin{aligned} &\int_U\nabla(\mathscr E_U A)\cdot\nabla(\mathscr E_U B)\,dx -\langle J\mathscr E_U A,J\mathscr E_U B\rangle_{L^2(\nu_U)}\\ &\hspace{1em}=\int_U\nabla A\cdot\nabla B\,dx -\langle JA,\mathscr R(T_U)JB\rangle_{L^2(\nu_U)}. \end{aligned}(12)✦

Let Ω\Omega have the closed disk conformal chart above, assume μ>0\mu>0, and take w,zw,z smooth on Ω¯\overline\Omega. Their scalar solutions satisfy (13), with RR formed from the conformal potential ν\nu. ∫Ω∇uw⋅∇uzdx−μ∫Ωuwuzdx=12∫S1×S1N0(s,t)(hw(s)−hw(t))(hz(s)−hz(t))dsdt−∫S1×S1R(s,t)hw(s)hz(t)dsdt.\begin{aligned} &\int_\Omega\nabla u_w\cdot\nabla u_z\,dx -\mu\int_\Omega u_wu_z\,dx\\ &=\frac12\int_{S^1\times S^1}N_0(s,t) \bigl(h_w(s)-h_w(t)\bigr)\bigl(h_z(s)-h_z(t)\bigr)\,ds\,dt\\ &\quad-\int_{S^1\times S^1}R(s,t)h_w(s)h_z(t)\,ds\,dt. \end{aligned}(13)✦

For data h,k:S1→Eh,k:S^1\to E in a real inner product space EE, a real kernel R̃:S1×S1→ℝ\widetilde R:S^1\times S^1\to\mathbb R, and a real coefficient b:S1→ℝb:S^1\to\mathbb R, the abstract boundary form ℬR̃,b\mathcal B_{\widetilde R,b} on these data is defined by (14). ℬR̃,b(h,k)=12∫S1×S1N0(s,t)⟨h(s)−h(t),k(s)−k(t)⟩Edsdt−∫S1×S1R̃(s,t)⟨h(s),k(t)⟩Edsdt+∫S1b(s)⟨h(s),k(s)⟩Eds.\begin{aligned} \mathcal B_{\widetilde R,b}(h,k) &=\frac12\int_{S^1\times S^1}N_0(s,t) \langle h(s)-h(t),k(s)-k(t)\rangle_E\,ds\,dt\\ &\quad-\int_{S^1\times S^1}\widetilde R(s,t) \langle h(s),k(t)\rangle_E\,ds\,dt +\int_{S^1}b(s)\langle h(s),k(s)\rangle_E\,ds. \end{aligned}(14)✦

In the paper, the result reads (15): for the original potential q=μ|Φ′|2dxq=\mu|\Phi'|^2\,dx and tangent Lipschitz data h,k:S1→ℝ2h,k:S^1\to\mathbb R^2, with WhW_h the disk extension, QQ the divergence-curl energy, and boundary curvature κ=det⁡(τ,∂ττ)\kappa=\det(\tau,\partial_\tau\tau) for the unit tangent τ\tau, Xh=Wh∘Φ−1,E(h,k)=Q(Xh,Xk)−μ⟨Xh,Xk⟩L2(Ω),E(h,k)=12∫S1×S1N0(s,t)(h(s)−h(t))⋅(k(s)−k(t))dsdt−∫S1×S1R(s,t)h(s)⋅k(t)dsdt+∫S1|Φ′|(κ∘Φ)h⋅kds.\begin{aligned} X_h&=W_h\circ\Phi^{-1}, &E(h,k)&=Q(X_h,X_k)-\mu\langle X_h,X_k\rangle_{L^2(\Omega)},\\ E(h,k)&=\frac12\int_{S^1\times S^1}N_0(s,t) \bigl(h(s)-h(t)\bigr)\cdot\bigl(k(s)-k(t)\bigr)\,ds\,dt\\ &\quad-\int_{S^1\times S^1}R(s,t)h(s)\cdot k(t)\,ds\,dt +\int_{S^1}|\Phi'|(\kappa\circ\Phi)h\cdot k\,ds. \end{aligned}(15)

What Is Checked

Identity (12) is stated for weakly harmonic Sobolev functions on a bounded open set, a finitely dominated potential measure, and a weighted Dirichlet Green operator of norm less than one. Identity (13) uses a nonempty bounded smooth domain with a closed disk conformal chart, positive first Neumann value, and smooth scalar data on its closure. The boundary form (14) is stated for data in a real inner product space.

The Dirichlet energy form on H01(U)H_0^1(U) is coercive when UU is bounded.✦ The conformal density bound CΦC_\Phi is finite✦ and, for μ≥0\mu\geq0, satisfies ν≤CΦdx|D\nu\leq C_\Phi\,dx|_D✦; the potential measure is finite on a bounded domain.✦ The Green rows belong to L2(D,dx)L^2(D,dx) for almost every xx✦, and their L2L^2 norms are uniformly bounded for x∈Dx\in D.

Smooth data ww on Ω¯\overline\Omega have a Lipschitz pulled boundary trace hwh_w.✦ Their pullbacks w∘Fw\circ F belong to H1(D)H^1(D) with their classical gradients as weak gradients.✦