OctoInformBeta

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

Lemma 3.2: The defect and nonlinear energy imbalance for a fixed radial cutoff

A relative bound for virial localization.

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 radial cutoff gives two localized nonlinear energies for the Hénon-Lane-Emden system. Their imbalance is coupled to a defect in the Newton interaction of the two weighted sources.

Let n≥2n\geq2 be an integer, let p,q>0p,q>0, and let A,B∈ℝA,B\in\mathbb R. A positive solution (u,v)(u,v) on the Euclidean space ℝn\mathbb R^n is a pair of continuous, strictly positive real functions, both of class C2C^2 off the origin, satisfying (1) there: −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)✦

The positive-part power (t)+j(t)_+^j of a real number tt, for a nonnegative integer exponent jj, is the real number obtained by raising its positive part to the power jj, with the notation in (2); the auxiliary cutoff integers m,k,Nm,k,N depend only on the dimension: t+=max⁡{t,0},m=n+3,k=n(m+1),N=k+n+2.t_+=\max\{t,0\},\qquad m=n+3,\qquad k=n(m+1),\qquad N=k+n+2.(2) The radial cutoff profile η0:ℝ→ℝ\eta_0:\mathbb R\to\mathbb R, the cutoff η:ℝn→ℝ\eta:\mathbb R^n\to\mathbb R, and its vector field XX and radial derivative weight ww are the functions in (3): η0(r)=(1−(r−1)+3)+N,η(x)=η0(|x|),X(x)=η(x)x,w(x)=−∇η(x)⋅x.\begin{aligned} \eta_0(r)&=\bigl(1-(r-1)_+^3\bigr)_+^N,& \eta(x)&=\eta_0(|x|),\\ X(x)&=\eta(x)x,& w(x)&=-\nabla\eta(x)\cdot x. \end{aligned}(3)

The weighted power sources f,g:ℝn→ℝf,g:\mathbb R^n\to\mathbb R of a positive solution are defined by (4): f(x)=|x|Bu(x)q,g(x)=|x|Av(x)p(x≠0),f(0)=0,g(0)=0.\begin{aligned} f(x)&=|x|^B u(x)^q,& g(x)&=|x|^A v(x)^p &&(x\ne0),\\ f(0)&=0,&g(0)&=0. \end{aligned}(4)✦ For a real weight χ\chi on ℝn\mathbb R^n, the weighted nonlinear energies Eχ(u;q,B)E_\chi(u;q,B) and Eχ(v;p,A)E_\chi(v;p,A) of a positive solution (u,v)(u,v) are the real integrals in (5); Ef,EgE_f,E_g denote the energies with weight η\eta, and EE their sum: Eχ(u;q,B)=∫ℝnχ(x)f(x)u(x)dx,Eχ(v;p,A)=∫ℝnχ(x)g(x)v(x)dx,Ef=Eη(u;q,B),Eg=Eη(v;p,A),E=Ef+Eg.\begin{aligned} E_\chi(u;q,B)&=\int_{\mathbb R^n}\chi(x)f(x)u(x)\,dx,& E_\chi(v;p,A)&=\int_{\mathbb R^n}\chi(x)g(x)v(x)\,dx,\\ E_f&=E_\eta(u;q,B),&E_g&=E_\eta(v;p,A),\qquad E=E_f+E_g. \end{aligned}(5)

Let σn−1\sigma_{n-1} be the unnormalized surface measure of the entire unit sphere induced by Lebesgue measure.✦ The Newton normalization dnd_n is the real number and the Newton kernel Kn:ℝn×ℝn→ℝK_n:\mathbb R^n\times\mathbb R^n\to\mathbb R is the function given in (6): dn=(n−2)σn−1,Kn(x,y)=dn−1|x−y|2−n.d_n=(n-2)\sigma_{n-1},\qquad K_n(x,y)=d_n^{-1}|x-y|^{2-n}.(6) Reciprocals at zero are assigned zero, and real powers at zero have value zero except for 00=10^0=1. The Newton interaction measure π=πf,g\pi=\pi_{f,g} on ℝn×ℝn\mathbb R^n\times\mathbb R^n is the measure with the nonnegative extended density in (7): 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.(7)✦

The Virial Defect

The virial flux factor Φ\Phi and radial alignment factor DD are the real functions on pairs of points defined by (8), with each quotient assigned zero when its denominator vanishes: Φ(x,y)=(x−y)⋅(X(x)−X(y))|x−y|2,D(x,y)=(x⋅(x−y))2|x|2|x−y|2.\Phi(x,y)=\frac{(x-y)\cdot(X(x)-X(y))}{|x-y|^2},\qquad D(x,y)=\frac{\bigl(x\cdot(x-y)\bigr)^2}{|x|^2|x-y|^2}.(8) The local virial error ℓ:ℝn×ℝn→ℝ\ell:\mathbb R^n\times\mathbb R^n\to\mathbb R is the nonnegative function in (9), where ℛ\mathscr R is the defect integrand: ℛ(x,y)=Φ(x,y)−η(x)+w(x)D(x,y),ℓ(x,y)=|ℛ(x,y)|+|η(x)−η(y)|.\begin{aligned} \mathscr R(x,y)&=\Phi(x,y)-\eta(x)+w(x)D(x,y),\\ \ell(x,y)&=|\mathscr R(x,y)|+|\eta(x)-\eta(y)|. \end{aligned}(9)✦ The integrated virial defect Q(f,g)Q(f,g) is the real interaction integral in (10): Q(f,g)=∫ℝn×ℝnℛ(x,y)dπf,g(x,y).Q(f,g)=\int_{\mathbb R^n\times\mathbb R^n}\mathscr R(x,y)\,d\pi_{f,g}(x,y).(10)✦

The Universal Constants

With B1B_1 denoting the unit ball centered at the origin, the unit ball weight mass Wn(t)W_n(t), for a real exponent tt, and the ball kernel lower bound constant cnc_n are the real numbers in (11): Wn(t)=∫B1|x|tdx,cn=22−ndn.W_n(t)=\int_{B_1}|x|^t\,dx,\qquad c_n=\frac{2^{2-n}}{d_n}.(11) The scaling exponents α,β\alpha,\beta of the system (1) are the real numbers given by (12): α=2+A+p(2+B)pq−1,β=2+B+q(2+A)pq−1.\alpha=\frac{2+A+p(2+B)}{pq-1},\qquad \beta=\frac{2+B+q(2+A)}{pq-1}.(12) The source mass growth constants Cg,CfC_g,C_f are the real numbers defined by (13): Cg=[(cnWn(A))(cnWn(B))p]1/(1−pq)cn,Cf=[(cnWn(B))(cnWn(A))q]1/(1−pq)cn.C_g=\frac{\bigl[(c_nW_n(A))(c_nW_n(B))^p\bigr]^{1/(1-pq)}}{c_n}, \qquad C_f=\frac{\bigl[(c_nW_n(B))(c_nW_n(A))^q\bigr]^{1/(1-pq)}}{c_n}.(13)

For real C,κ,sC,\kappa,s, the dyadic tail bound constant T(C,κ,s)T(C,\kappa,s) is the real number in (14); its specializations Ug,UfU_g,U_f are the uniform tail bound constants: T(C,κ,s)=2κC2s8s−κ1−2s−κ,Ug=T(Cg,n−2,n−2−α),Uf=T(Cf,n−2,n−2−β).\begin{aligned} T(C,\kappa,s)&=2^\kappa\frac{C\,2^s8^{s-\kappa}}{1-2^{s-\kappa}},\\ U_g&=T(C_g,n-2,n-2-\alpha),\qquad U_f=T(C_f,n-2,n-2-\beta). \end{aligned}(14) The truncated potential bound constants Pg,PfP_g,P_f are the real numbers defined by (15): Pg=max⁡{dn−1(Cg8n−2−α+Ug),0},Pf=max⁡{dn−1(Cf8n−2−β+Uf),0}.P_g=\max\{d_n^{-1}(C_g8^{n-2-\alpha}+U_g),0\},\qquad P_f=\max\{d_n^{-1}(C_f8^{n-2-\beta}+U_f),0\}.(15)

The first and second cutoff derivative constants d1,n,d2,nd_{1,n},d_{2,n}, the near interaction error constant cnearc_{\mathrm{near}}, and the far localization error coefficient hnh_n are the dimension-dependent numbers in (16): d1,n=3N3N−1,d2,n=(6N+9N(N−1))3N−2,cnear=(2d2,n+4d1,n)2N,hn=2⋅3N+2d1,n.\begin{aligned} d_{1,n}&=3N3^{N-1},& d_{2,n}&=\bigl(6N+9N(N-1)\bigr)3^{N-2},\\ c_{\mathrm{near}}&=(2d_{2,n}+4d_{1,n})2^N,& h_n&=2\cdot3^N+2d_{1,n}. \end{aligned}(16) For δ>0\delta>0, the far interaction bound constants Fg(δ),Ff(δ)F_g(\delta),F_f(\delta) and localization remainder constant L(δ)L(\delta) are the real numbers in (17): Fg(δ)=Pgδ2−nmax⁡{Cf2n−2−β,0},Ff(δ)=Pfδ2−nmax⁡{Cg2n−2−α,0},L(δ)=hn(1+δ−1)(Fg(δ)+Ff(δ)).\begin{aligned} F_g(\delta)&=P_g\delta^{2-n}\max\{C_f2^{n-2-\beta},0\},\\ F_f(\delta)&=P_f\delta^{2-n}\max\{C_g2^{n-2-\alpha},0\},\\ L(\delta)&=h_n(1+\delta^{-1})\bigl(F_g(\delta)+F_f(\delta)\bigr). \end{aligned}(17)

The Claim

Theorem 1. Fix an integer n≥2n\geq2, positive powers p,qp,q, and real weights A,BA,B. For every positive solution (u,v)(u,v) of (1) and every 0<δ≤140<\delta\leq\tfrac14, the local error bound (18) holds: ℓ∈L1(πf,g),∫ℝn×ℝnℓdπf,g≤cnearδE+L(δ).\ell\in L^1(\pi_{f,g}),\qquad \int_{\mathbb R^n\times\mathbb R^n}\ell\,d\pi_{f,g} \leq c_{\mathrm{near}}\delta E+L(\delta).(18)✦ Given ε>0\varepsilon>0, every positive solution satisfies the relative localization bound (19) with a constant Cε≥0C_\varepsilon\geq0 depending only on n,p,q,A,B,εn,p,q,A,B,\varepsilon: |Q(f,g)|+|Ef−Eg|≤εE+Cε.|Q(f,g)|+|E_f-E_g|\leq\varepsilon E+C_\varepsilon.(19)✦

In the paper, the result reads: under its reduced solution hypotheses and fixed cutoffs, every ε>0\varepsilon>0 admits a solution-independent constant CεC_\varepsilon such that |Ef−Eg|+∫|ℛ(x,y)|dπ(x,y)≤εE+Cε|E_f-E_g|+\int|\mathscr R(x,y)|\,d\pi(x,y)\leq\varepsilon E+C_\varepsilon.

What Is Checked

The two estimates concern positive solutions as defined by (1), with n≥2n\geq2, p,q>0p,q>0 and real A,BA,B, and the fixed cutoff and Newton interaction constructions in (3) and (7).