OctoInformBeta

OpenAI preprint, September 2026 · Section 3. A localized virial identity

Section 3 interaction and cutoffs: Marginals, alignment factors and radial cutoffs for positive solutions

Interaction measures and localized energy weights.

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 Newton interaction couples the two weighted sources of the Hénon-Lane-Emden system, while a spatial weight specifies the nonlinear energy being localized. Work in Euclidean coordinate space ℝn\mathbb R^n with Lebesgue measure. For real parameters p,q,A,Bp,q,A,B, a positive solution (u,v)(u,v) is a pair of continuous, strictly positive real-valued functions on ℝn\mathbb R^n, of class C2C^2 off the origin, satisfying (1) at every nonzero point: −Δu(x)=|x|Av(x)p,−Δv(x)=|x|Bu(x)q,x≠0.-\Delta u(x)=|x|^A v(x)^p, \qquad -\Delta v(x)=|x|^B u(x)^q, \qquad x\ne0.(1)✦

For a real-valued function uu on ℝn\mathbb R^n and real parameters q,Bq,B, the weighted power source fu,q,B:ℝn→ℝf_{u,q,B}:\mathbb R^n\to\mathbb R is the function in (2); ff and gg denote the sources associated with a solution: fu,q,B(x)={|x|Bu(x)q,x≠0,0,x=0,f=fu,q,B,g=fv,p,A.f_{u,q,B}(x)= \begin{cases} |x|^B u(x)^q,&x\ne0,\\ 0,&x=0, \end{cases} \qquad f=f_{u,q,B},\qquad g=f_{v,p,A}.(2)✦ All powers are real powers, with 00=10^0=1 and 0t=00^t=0 for t≠0t\ne0; reciprocals and quotients with zero denominator are assigned zero.

Let σn\sigma_n be the area of the unit sphere for the unnormalized surface measure induced by Lebesgue measure.✦ The Newton normalization cnc_n is the real constant and the Newton kernel KnK_n on ℝn×ℝn\mathbb R^n\times\mathbb R^n is the real-valued function defined by (3), including on the diagonal: cn=(n−2)σn,Kn(x,y)=cn−1|x−y|2−n.c_n=(n-2)\sigma_n, \qquad K_n(x,y)=c_n^{-1}|x-y|^{2-n}.(3)

For an integer ℓ≥0\ell\ge0, the positive-part power Pℓ:ℝ→ℝP_\ell:\mathbb R\to\mathbb R is the function defined for r∈ℝr\in\mathbb R by (4): r+=max⁡{r,0},Pℓ(r)=(r+)ℓ.r_+=\max\{r,0\},\qquad P_\ell(r)=(r_+)^\ell.(4)✦

The Claim

Definition 1 (Interaction and localization). For real functions f,gf,g on ℝn\mathbb R^n, the Newton interaction πf,g\pi_{f,g} is the nonnegative measure on ℝn×ℝn\mathbb R^n\times\mathbb R^n specified by (5): dπf,g(x,y)=f(x)+g(y)+Kn(x,y)+dxdy.d\pi_{f,g}(x,y)=f(x)_+g(y)_+K_n(x,y)_+\,dx\,dy.(5)✦

Let EE be a real inner product space. For a unit vector e∈Ee\in E, the directional alignment factor De:E×E→ℝD_e:E\times E\to\mathbb R is given by (6); the radial alignment factor D:E×E→ℝD:E\times E\to\mathbb R is given by (7): De(x,y)=⟨e,x−y⟩2|x−y|2.D_e(x,y)=\frac{\langle e,x-y\rangle^2}{|x-y|^2}.(6)✦ D(x,y)=⟨x,x−y⟩2|x|2|x−y|2.D(x,y)=\frac{\langle x,x-y\rangle^2}{|x|^2|x-y|^2}.(7)✦

The boundary gap profile δ:ℝ→ℝ\delta:\mathbb R\to\mathbb R and boundary gap d:E→ℝd:E\to\mathbb R are defined by (8) and (9): δ(r)=min⁡{1,(2−r)+}.\delta(r)=\min\{1,(2-r)_+\}.(8)✦ d(x)=δ(|x|).d(x)=\delta(|x|).(9)✦

For a nonnegative integer nn, the auxiliary integers m,km,k and cutoff exponent NN are prescribed by (10), (11) and (12), respectively: m=n+3.m=n+3.(10)✦ k=n(m+1).k=n(m+1).(11)✦ N=k+n+2.N=k+n+2.(12)✦

The radial cutoff profile ϑn:ℝ→ℝ\vartheta_n:\mathbb R\to\mathbb R and compact cutoff η:E→ℝ\eta:E\to\mathbb R are the functions in (13) and (14): ϑn(r)=PN(1−P3(r−1)).\vartheta_n(r)=P_N\bigl(1-P_3(r-1)\bigr).(13)✦ η(x)=ϑn(|x|).\eta(x)=\vartheta_n(|x|).(14)✦

The radial cutoff weight w:E→ℝw:E\to\mathbb R and cutoff vector field X:E→EX:E\to E are given by (15) and (16), with dη(x)[x]\mathrm d\eta(x)[x] denoting the Fréchet derivative applied to xx: w(x)=−dη(x)[x].w(x)=-\mathrm d\eta(x)[x].(15)✦ X(x)=η(x)x.X(x)=\eta(x)x.(16)✦ The localized flux factor Φ:E×E→ℝ\Phi:E\times E\to\mathbb R associated with XX is (17): Φ(x,y)=⟨x−y,X(x)−X(y)⟩|x−y|2.\Phi(x,y)=\frac{\langle x-y,X(x)-X(y)\rangle}{|x-y|^2}.(17)✦

For real functions χ,u\chi,u on ℝn\mathbb R^n and real parameters q,Bq,B, the weighted nonlinear energy ℰχ(u;q,B)\mathcal E_\chi(u;q,B) is the real-valued integral (18): ℰχ(u;q,B)=∫ℝnχ(x)fu,q,B(x)u(x)dx.\mathcal E_\chi(u;q,B) =\int_{\mathbb R^n}\chi(x)f_{u,q,B}(x)u(x)\,dx.(18)✦

Theorem 1 (Interaction marginals). Let n≥2n\ge2 and let (u,v)(u,v) be a positive solution with real parameters p,q,A,Bp,q,A,B. Let F:ℝn→[0,∞]F:\mathbb R^n\to[0,\infty] be measurable. If q>0q>0, the left marginal identity is (19), as an equality of nonnegative extended integrals: ∫ℝn×ℝnF(x)dπf,g(x,y)=∫ℝnf(x)u(x)F(x)dx.\int_{\mathbb R^n\times\mathbb R^n}F(x)\,d\pi_{f,g}(x,y) =\int_{\mathbb R^n}f(x)u(x)F(x)\,dx.(19)✦ If p>0p>0, the right marginal identity is (20), again for nonnegative extended integrals: ∫ℝn×ℝnF(y)dπf,g(x,y)=∫ℝng(y)v(y)F(y)dy.\int_{\mathbb R^n\times\mathbb R^n}F(y)\,d\pi_{f,g}(x,y) =\int_{\mathbb R^n}g(y)v(y)F(y)\,dy.(20)✦

In the paper, the result reads: dπ(x,y)=f(x)g(y)Kn(x,y)dxdyd\pi(x,y)=f(x)g(y)K_n(x,y)\,dx\,dy, with marginals f(x)u(x)dxf(x)u(x)\,dx and g(y)v(y)dyg(y)v(y)\,dy; for |e|=1|e|=1, De(x,y)=((x−y)⋅e)2/|x−y|2D_e(x,y)=((x-y)\cdot e)^2/|x-y|^2 and D(x,y)=Dx/|x|(x,y)D(x,y)=D_{x/|x|}(x,y) off x=0x=0; set d(x)=min⁡{1,(2−|x|)+}d(x)=\min\{1,(2-|x|)_+\}, m=n+3m=n+3, k=n(m+1)k=n(m+1), N=k+n+2N=k+n+2, η(x)=[(1−(|x|−1)+3)+]N\eta(x)=[(1-(|x|-1)_+^3)_+]^N, w(x)=−x⋅∇η(x)w(x)=-x\cdot\nabla\eta(x) and X(x)=xη(x)X(x)=x\eta(x), and define Ef=∫ηfuE_f=\int\eta fu, Eg=∫ηgvE_g=\int\eta gv, E=Ef+EgE=E_f+E_g, Wf=∫wfuW_f=\int wfu, Wg=∫wgvW_g=\int wgv and Φ(x,y)=(X(x)−X(y))⋅(x−y)/|x−y|2\Phi(x,y)=(X(x)-X(y))\cdot(x-y)/|x-y|^2, these energies being finite for each solution in the reduced range.

What Is Checked

The marginal identities are stated over positive solutions on ℝn\mathbb R^n with n≥2n\ge2 and nonnegative measurable tests, requiring q>0q>0 for the left identity and p>0p>0 for the right; the geometric definitions use a real inner product space and unit directions, the cutoff definitions allow every nonnegative integer nn, and the weighted energy is defined for real-valued weights and functions.