OctoInformBeta

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

Lemma 4.1: The weighted Dirichlet inequality controls its operator norm

A symmetric Green operator for disk measures.

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

A disk measure specifies the square-integrable data on which the weighted Green operator acts. The energy comparison concerns the Dirichlet energy of functions with zero boundary values. Let D={x∈ℝ2:|x|<1}D=\{x\in\mathbb R^2:|x|<1\} be the open unit disk, and write dAdA for area measure on DD.✦ All function spaces are real. A measure ν\nu on DD has a finite density bound when, for some finite C≥0C\geq0, it satisfies (1): 0≤ν≤CdA.0\leq\nu\leq C\,dA.(1)

The real linear isometry ι:ℂ→ℝ2\iota:\mathbb C\to\mathbb R^2 sends zz to (Re⁡z,Im⁡z)(\operatorname{Re}z,\operatorname{Im}z).✦ The complex disk Dirichlet Green kernel GℂG_{\mathbb C} on ι−1D\iota^{-1}D and its planar version GG on DD are defined by (2), with diagonal value zero: Gℂ(z,w)=14πlog⁡(1+(1−|z|2)(1−|w|2)|z−w|2),z≠w,G(x,y)=Gℂ(ι−1x,ι−1y).\begin{aligned} G_{\mathbb C}(z,w)&=\frac{1}{4\pi} \log\left(1+\frac{(1-|z|^2)(1-|w|^2)}{|z-w|^2}\right), && z\ne w,\\ G(x,y)&=G_{\mathbb C}(\iota^{-1}x,\iota^{-1}y). \end{aligned}(2) The area-measure Green integral operator AA is the integral operator on L2(D,dA)L^2(D,dA) given by (3): (Af)(x)=∫DG(x,y)f(y)dA(y).(Af)(x)=\int_D G(x,y)f(y)\,dA(y).(3)✦

Sobolev Functions and Energy

A test function φ∈Cc∞(D)\varphi\in C_c^\infty(D) is a smooth real function on the plane with compact closed support contained in DD.✦ For a test function φ∈Cc∞(D)\varphi\in C_c^\infty(D) and a direction e∈ℝ2e\in\mathbb R^2, the test derivative ∂eφ\partial_e\varphi is the directional derivative regarded as a scalar L2(D,dA)L^2(D,dA) class, and the test vector field φe\varphi e is the product regarded as a vector-valued L2(D,dA)L^2(D,dA) class.

The ambient space ℋ\mathcal H of scalar-vector pairs has the Euclidean product norm in (4): ℋ=L2(D,dA)×L2(D,dA;ℝ2),∥(v,g)∥ℋ2=∥v∥L2(D,dA)2+∥g∥L2(D,dA;ℝ2)2.\begin{aligned} \mathcal H&=L^2(D,dA)\times L^2(D,dA;\mathbb R^2),\\ \|(v,g)\|_{\mathcal H}^2 &=\|v\|_{L^2(D,dA)}^2+\|g\|_{L^2(D,dA;\mathbb R^2)}^2. \end{aligned}(4)✦

A vector field gg is the weak gradient of a real function vv when both are square integrable and satisfy (5) for every test function φ\varphi and direction ee: ∫Dv∂eφdA=−∫D(g⋅e)φdA.\int_D v\,\partial_e\varphi\,dA =-\int_D (g\cdot e)\varphi\,dA.(5)✦ The Sobolev space H1(D)H^1(D) is the weak gradient graph in ℋ\mathcal H✦; the Sobolev class of vv with weak gradient gg is the corresponding pair of L2L^2 classes.✦ The Sobolev value map val\operatorname{val} and Sobolev gradient ∇\nabla on H1(D)H^1(D) are the continuous linear maps selecting the scalar and vector components of a Sobolev class, respectively. The Sobolev norm is the norm inherited from ℋ\mathcal H in (4).

The test-function embedding sends φ∈Cc∞(D)\varphi\in C_c^\infty(D) to its Sobolev class with its classical gradient.✦ The Dirichlet Sobolev space H01(D)H_0^1(D) is the closed subspace of H1(D)H^1(D) obtained by taking the closure of the Sobolev classes of test functions in the Sobolev norm. The Dirichlet value and gradient maps on H01(D)H_0^1(D) are the restrictions of val\operatorname{val} and ∇\nabla to H01(D)H_0^1(D).

Weighted Values and the Green Operator

The weighted inclusion J:L2(D,dA)→L2(D,ν)J:L^2(D,dA)\to L^2(D,\nu) is the continuous linear map sending an area-measure class to its ν\nu-class under (1).✦ The weighted Sobolev value map val⁡ν:H01(D)→L2(D,ν)\operatorname{val}_{\nu}:H_0^1(D)\to L^2(D,\nu) is the continuous linear map in (6): val⁡νu=J(val⁡u).\operatorname{val}_{\nu}u=J(\operatorname{val}u).(6)✦ The weighted Green operator TT on L2(D,ν)L^2(D,\nu) is the continuous linear operator defined in (7), where J*J^* is the Hilbert-space adjoint of JJ: T=JAJ*.T=JAJ^*.(7)✦

The Claim

Theorem 1. Let ν\nu be a disk measure satisfying (1), and let TT be the weighted Green operator (7). For all f,g∈L2(D,ν)f,g\in L^2(D,\nu), ⟨Tf,g⟩L2(D,ν)=⟨f,Tg⟩L2(D,ν).\langle Tf,g\rangle_{L^2(D,\nu)} =\langle f,Tg\rangle_{L^2(D,\nu)}.(8)✦ Let d≥0d\geq0, and suppose that the weighted Dirichlet inequality (9) holds: ∥val⁡νu∥L2(D,ν)2≤d∥∇u∥L2(D,dA;ℝ2)2for every u∈H01(D).\|\operatorname{val}_{\nu}u\|_{L^2(D,\nu)}^2 \leq d\,\|\nabla u\|_{L^2(D,dA;\mathbb R^2)}^2 \qquad\text{for every }u\in H_0^1(D).(9) Then ∥T∥L2(D,ν)→L2(D,ν)≤d.\|T\|_{L^2(D,\nu)\to L^2(D,\nu)}\leq d.(10)✦

In the paper, the result reads: for a disk measure qq with nonnegative bounded density satisfying ∫D|u|2dq≤d∫D|∇u|2dA\int_D|u|^2\,dq\leq d\int_D|\nabla u|^2\,dA for every u∈H01(D)u\in H_0^1(D), with 0<d<10<d<1, the operator Tf(x)=∫DG(x,y)f(y)dq(y)Tf(x)=\int_D G(x,y)f(y)\,dq(y) is bounded and self-adjoint on L2(q)L^2(q) with ∥T∥≤d<1\|T\|\leq d<1, and the same holds for every measure 0≤ρ≤q0\leq\rho\leq q.

What Is Checked

The assertions (10) and (8) concern the real Hilbert space L2(D,ν)L^2(D,\nu), the measure domination (1) and the weighted operator (7); the norm assertion additionally assumes (9) with d≥0d\geq0.

The Green rows have the square-integrability and uniform row bound in (11), for some B≥0B\geq0: G(x,⋅)∈L2(D,dA)for almost every x∈D,∥G(x,⋅)∥L2(D,dA)≤B(x∈D).\begin{aligned} G(x,\cdot)&\in L^2(D,dA) &&\text{for almost every }x\in D,\\ \|G(x,\cdot)\|_{L^2(D,dA)}&\leq B &&(x\in D). \end{aligned}(11) The L2L^2 kernel operator construction gives a square-integrable output (3) for every f∈L2(D,dA)f\in L^2(D,dA).✦