OctoInformBeta

OpenAI preprint, September 2026 · §3. Two local Coulomb branches

§3 (setting): the common central quadratic and stationary charts: Analytic stationary charts and exact local supporting values for K=20K=20

A common central quadratic for Coulomb branches.

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

In three-marginal Coulomb transport, a local branch selects a central point xx and an opposite outer point zz as functions of the positive outer point yy. Take E=ℝ3E=\mathbb R^3, with Euclidean norm |⋅||\cdot| and scalar product ⋅\cdot. The extended Coulomb cost c:E3→[0,+∞]c:E^3\to[0,+\infty] is given by (1). c(x,y,z)={1|x−y|+1|x−z|+1|y−z|,x,y,z pairwise distinct,+∞,at any collision.c(x,y,z)= \begin{cases} \displaystyle\frac{1}{|x-y|}+\frac{1}{|x-z|}+\frac{1}{|y-z|}, & x,y,z\text{ pairwise distinct},\\[4pt] +\infty, & \text{at any collision}. \end{cases}(1)

The local scalar interaction q:E→ℝq:E\to\mathbb R is the real-valued function in (2). q(ξ)={|ξ|−1,ξ≠0,0,ξ=0.q(\xi)= \begin{cases} |\xi|^{-1}, & \xi\ne0,\\ 0, & \xi=0. \end{cases}(2)✦ The Coulomb gradient field g:E→Eg:E\to E is the vector field in (3). g(ξ)={−ξ/|ξ|3,ξ≠0,0,ξ=0.g(\xi)= \begin{cases} -\xi/|\xi|^3, & \xi\ne0,\\ 0, & \xi=0. \end{cases}(3)✦

The Definition

Definition 1. Fix K=20K=20 and a vector a∈Ea\in E. The state potential Φa:E×(E×E)→ℝ\Phi_a:E\times(E\times E)\to\mathbb R is the scalar function defined by (4), with parameter yy and state (x,z)(x,z). Φa(y;(x,z))=q(x−y)+q(x−z)+q(z−y)−(52−10|x|2)−(54a⋅(z+a)−10|z+a|2).\begin{aligned} \Phi_a(y;(x,z))={}&q(x-y)+q(x-z)+q(z-y)\\ &-\left(\frac52-10|x|^2\right) -\left(\frac54 a\cdot(z+a)-10|z+a|^2\right). \end{aligned}(4)✦ The joint scalar potential ℱa:E×(E×E)→ℝ\mathcal F_a:E\times(E\times E)\to\mathbb R is the function in (5). ℱa(y,(x,z))=Φa(y;(x,z)).\mathcal F_a\bigl(y,(x,z)\bigr)=\Phi_a(y;(x,z)).(5)✦ The stationarity map Ga:E×(E×E)→E×EG_a:E\times(E\times E)\to E\times E is the vector map defined by (6). Ga(y,(x,z))=(g(x−y)+g(x−z)+20x,g(z−x)+g(z−y)−54a+20(z+a)).G_a\bigl(y,(x,z)\bigr)= \left( \begin{aligned} &g(x-y)+g(x-z)+20x,\\ &g(z-x)+g(z-y)-\tfrac54a+20(z+a) \end{aligned} \right).(6)✦

Let P,SP,S be real normed spaces, let G:P×(S×S)→S×SG:P\times(S\times S)\to S\times S, and fix p∈Pp\in P and s∈S×Ss\in S\times S. Analytic stationary charts for GG at (p,s)(p,s) are maps X,Z:P→SX,Z:P\to S whose restrictions to a common open source U⊂PU\subset P are real analytic diffeomorphisms onto open targets in SS, with real analytic inverses, together with an open state set D⊂S×SD\subset S\times S satisfying (7). p∈U,s∈D,(X(p),Z(p))=s,(X(y),Z(y))∈D(y∈U),G(y,(X(y),Z(y)))=0(y∈U),G(y,σ)=0⇔σ=(X(y),Z(y))(y∈U,σ∈D).\begin{gathered} p\in U,\qquad s\in D,\qquad (X(p),Z(p))=s,\\ (X(y),Z(y))\in D\qquad(y\in U),\\ G\bigl(y,(X(y),Z(y))\bigr)=0\qquad(y\in U),\\ G(y,\sigma)=0\quad\Longleftrightarrow\quad \sigma=(X(y),Z(y)) \qquad(y\in U,\ \sigma\in D). \end{gathered}(7)✦

Let F:E×(E×E)→ℝF:E\times(E\times E)\to\mathbb R and G:E×(E×E)→E×EG:E\times(E\times E)\to E\times E, and fix p∈Ep\in E and s∈E×Es\in E\times E. Local supporting charts for (F,G)(F,G) at (p,s)(p,s) are analytic stationary charts X,ZX,Z for GG at (p,s)(p,s), with positive radii r,Rr,R and a function v:E→ℝv:E\to\mathbb R real analytic on their source UU, satisfying (8); BB denotes an open ball in the corresponding metric space. U=B(p,r),D=B(s,R),r>0,R>0,v(y)=F(y,(X(y),Z(y)))(y∈E),v(p)=F(p,s),v(y)≤F(y,σ)(y∈U,σ∈D),F(y,σ)=v(y)⇔σ=(X(y),Z(y))(y∈U,σ∈D).\begin{gathered} U=B(p,r),\qquad D=B(s,R),\qquad r>0,\quad R>0,\\ v(y)=F\bigl(y,(X(y),Z(y))\bigr)\qquad(y\in E), \qquad v(p)=F(p,s),\\ v(y)\le F(y,\sigma)\qquad(y\in U,\ \sigma\in D),\\ F(y,\sigma)=v(y)\quad\Longleftrightarrow\quad \sigma=(X(y),Z(y)) \qquad(y\in U,\ \sigma\in D). \end{gathered}(8)✦

In the paper, the result reads: for a1=e1a_1=e_1 and a2=e2a_2=e_2, where e1,e2e_1,e_2 are the first two standard basis vectors of EE, the positive outer point yy parametrizes the stationary pair (Xk(y),Zk(y))(X_k(y),Z_k(y)) in (x,z)(x,z) for Fk(x,y,z)=c(x,y,z)−u0(x)−vk(z)F_k(x,y,z)=c(x,y,z)-u_0(x)-v_k(z), where u0(x)=52−K2|x|2u_0(x)=\frac52-\frac K2|x|^2 and vk(z)=54ak⋅(z+ak)−K2|z+ak|2v_k(z)=\frac54 a_k\cdot(z+a_k)-\frac K2|z+a_k|^2.

What Is Checked

The scalar potential and stationarity map are defined for every y∈Ey\in E and (x,z)∈E×E(x,z)\in E\times E; the chart definitions are stated over open parameter and state neighborhoods, with open balls in the local supporting condition.