OctoInformBeta

Nonattainment of the three-marginal Coulomb Monge problem

For a Borel probability measure μ\mu on ℝ3\mathbb R^3, let Π(μ,μ,μ)\Pi(\mu,\mu,\mu) denote the couplings with all three coordinate marginals equal to μ\mu. For x1,x2,x3∈ℝ3x_1,x_2,x_3\in\mathbb R^3, the Coulomb cost is c(x1,x2,x3)=∑1≤i<j≤3‖xi−xj‖2−1c(x_1,x_2,x_3)=\sum_{1\le i<j\le3}\lVert x_i-x_j\rVert_2^{-1}, with value +∞+\infty when two coordinates coincide. The Kantorovich value is E(μ)=inf⁡π∈Π(μ,μ,μ)∫cdπE(\mu)=\inf_{\pi\in\Pi(\mu,\mu,\mu)}\int c\,d\pi. A Monge plan is a coupling (Id,T2,T3)#μ(\mathrm{Id},T_2,T_3)_\#\mu for Borel maps T2,T3:ℝ3→ℝ3T_2,T_3:\mathbb R^3\to\mathbb R^3 satisfying (T2)#μ=(T3)#μ=μ(T_2)_\#\mu=(T_3)_\#\mu=\mu. The attainment question asks whether absolute continuity of μ\mu and E(μ)<∞E(\mu)<\infty guarantee an optimal Monge plan.

Colombo and Di Marino proved equality of the Coulomb Monge and Kantorovich infima for identical nonatomic marginals. For two particles, Cotar, Friesecke and Klüppelberg proved existence and uniqueness of an optimal Coulomb map for absolutely continuous marginals in every dimension. Colombo and Stra gave counterexamples to prescribed radial constructions for spherically symmetric Coulomb transport.

The preprint constructs a nonnegative probability density ρ\rho on ℝ3\mathbb R^3 with ρ,ρ∈Cc∞(ℝ3)\rho,\sqrt{\rho}\in C_c^\infty(\mathbb R^3). Write μ=ρdx\mu=\rho\,dx for its marginal. A finite-cost coupling attains E(μ)E(\mu). Every Monge plan has cost strictly greater than E(μ)E(\mu). For every integer d≥2d\ge2 and every s>0s>0, the inverse-power cost on (ℝd)3(\mathbb R^d)^3 is ∑1≤i<j≤3‖xi−xj‖2−s\sum_{1\le i<j\le3}\lVert x_i-x_j\rVert_2^{-s}. For these costs, the preprint establishes the same properties outside the Lean library.

Lean has checked every result of the Lean library. The Lean library is therefore the reference for the statements on these pages. Each result is stated as in the Lean library and linked to its corresponding Lean declaration. Where the preprint states a result differently, the Lean library’s statement appears with the preprint’s statement beside it. The Lean library’s main existence theorem has no additional hypotheses. The preprint is a guide only to the order and names of the results.

The Lean library: CoulombTransport

CoulombTransport is a Lean library: 14 items, in six groups. Each has a page written by OctoInform, listed below.

§1. The problem and the result 2 items
  1. §1 (setting): Coulomb couplings, preserving maps, and density regularity Coulomb transport and strict Monge nonattainment.
  2. Theorem 1.1 [Smooth Coulomb nonattainment] A Coulomb minimum without a Monge optimizer.
§2. An obstruction to deterministic selection 2 items
  1. §2 (setting): the five-component mixture Combining a measure with its branch images.
  2. Lemma 2.1 [No preserving deterministic branch selector] An obstruction to measure-preserving branch selection.
§3. Two local Coulomb branches 2 items
  1. §3 (setting): the common central quadratic and stationary charts A common central quadratic for Coulomb branches.
  2. Lemma 3.1 [Two local exact supporting families] Exact Coulomb supports near coordinate directions.
§4. A supporting inequality on five components 3 items
  1. §4 (setting): five local potentials and the symmetric graph union Five component potentials from two supporting charts.
  2. Proposition 4.1 [Global certificate on five components] Local Coulomb supports on five components.
  3. Lemma 4.2 [A common central ball for all four branches] Four branches share a central ball.
§5. The marginal and exclusion of every Monge optimizer 2 items
  1. §5 (setting): transported square amplitudes and the smooth marginal Disjoint smooth charts yield a smooth marginal.
  2. §5 (proof of Theorem 1.1): an optimal coupling and the selection obstruction Branch couplings and the obstruction to Monge optimality.
§6. Equality of the two infima 3 items
  1. §6 (setting): finite-cost approximation by preserving pairs Approximating the Kantorovich value by preserving pairs.
  2. Proposition 6.1 [Equality of infima with strict nonattainment] Compactly supported atomless marginals admit approximation from separated optimizers.
  3. Lemma 6.2 [Borel transport between absolutely continuous probabilities] Measurable transport from an atomless probability.

No item matches.