Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition 3.1.1 — (anti-)self-dual connections are absolute minimizers of Yang-Mills (box version)

Proved
UhlenbeckGauge.selfDual_absolute_minimizer

by Lucas · Oct 9, 2026 · Mathlib 0df444a (Lean v4.33.1)

differential-geometrygauge-theorymathematical-physicsyang-mills

Let a,b∈R4a,b\in\mathbb R^4a,b∈R4, Ω=[a,b]\Omega=[a,b]Ω=[a,b] the closed box, and let AAA, BBB be smooth SU(n)SU(n)SU(n)-connections on the trivial bundle with the same boundary data, A=BA=BA=B at every point of ∂Ω\partial\Omega∂Ω. Suppose the curvature of AAA is self-dual at every point of Ω\OmegaΩ, or anti-self-dual at every point of Ω\OmegaΩ. Then

YmΩ(A)=∫Ω∣FA∣2 dvol ≤ ∫Ω∣FB∣2 dvol=YmΩ(B).\mathrm{Ym}_\Omega(A)=\int_\Omega|F_A|^2\,d\mathrm{vol}\ \le\ \int_\Omega|F_B|^2\,d\mathrm{vol}=\mathrm{Ym}_\Omega(B).YmΩ​(A)=∫Ω​∣FA​∣2dvol ≤ ∫Ω​∣FB​∣2dvol=YmΩ​(B).

That is, solutions of the (anti-)self-dual Yang–Mills equations are absolute minimizers of the Yang–Mills functional among connections with the same boundary values.

Formalization Note The notes state the result within a topological class c2(P)=kc_2(P)=kc2​(P)=k on a general bundle over a four-manifold with boundary; here the bundle is the trivial SU(n)SU(n)SU(n)-bundle over a closed axis-parallel box in Euclidean R4\mathbb R^4R4, so there is a single topological class. Connections are smooth on all of R4\mathbb R^4R4.

Preamble
import Definitions.Def_uhlenbeck_gauge_box_defs

open UhlenbeckGauge
Formal statement
namespace UhlenbeckGauge

theorem selfDual_absolute_minimizer {n : ℕ} (a b : R4) (A B : Connection n)
    (hA : IsSUConnection A) (hB : IsSUConnection B)
    (hbdry : ∀ x, OnBoxBoundary a b x → A x = B x)
    (hdual : (∀ x ∈ Set.Icc a b, IsSelfDual (curvature A x)) ∨
      (∀ x ∈ Set.Icc a b, IsAntiSelfDual (curvature A x))) :
    yangMills a b A ≤ yangMills a b B := by sorry

end UhlenbeckGauge
Source
K. Uhlenbeck, *Equations of Gauge Theory*, lecture notes by L. Fredrickson (Emil Grosswald Lectures, Temple University, February 7-9, 2012), Chapter 3, Section 3.1, pp. 29-32, Proposition 3.1.1 (p. 31)
Read-back

What the Lean code literally says, in plain math · Aristotle (Harmonic) - same agent that drafted the statements; non-blind

Disclosure: this read-back is NOT blind. It was written by the same agent (Aristotle, by Harmonic) that drafted the Lean statements, at the explicit direction of the proposal owner, with full knowledge of the source notes and of the intended meaning. It is not independent testimony and must not be mistaken for an independent audit; a reviewer should compare the Lean code against the source directly.

For every natural number nnn, all a,b∈R4a,b\in\mathbb R^4a,b∈R4 (no ordering assumed; if aμ>bμa_\mu>b_\muaμ​>bμ​ for some μ\muμ the box is empty), and all maps A,B:R4→({0,1,2,3}→Matn(C))A,B:\mathbb R^4\to(\{0,1,2,3\}\to\mathrm{Mat}_n(\mathbb C))A,B:R4→({0,1,2,3}→Matn​(C)) such that

  1. AAA and BBB are SU(n)SU(n)SU(n)-connections: every matrix entry of every coefficient is C∞C^\inftyC∞ on all of R4\mathbb R^4R4, and every coefficient at every point is skew-Hermitian with trace 000;
  2. A(x)=B(x)A(x)=B(x)A(x)=B(x) (all four coefficients) at every point xxx of the boundary of the box, i.e. every xxx with aμ≤xμ≤bμa_\mu\le x_\mu\le b_\muaμ​≤xμ​≤bμ​ for all μ\muμ and xμ∈{aμ,bμ}x_\mu\in\{a_\mu,b_\mu\}xμ​∈{aμ​,bμ​} for at least one μ\muμ;
  3. either ⋆FA(x)=FA(x)\star F_A(x)=F_A(x)⋆FA​(x)=FA​(x) for every xxx in the closed box [a,b][a,b][a,b], or ⋆FA(x)=−FA(x)\star F_A(x)=-F_A(x)⋆FA​(x)=−FA​(x) for every x∈[a,b]x\in[a,b]x∈[a,b],

the inequality

∫[a,b]∣FA(x)∣2 dx ≤ ∫[a,b]∣FB(x)∣2 dx\int_{[a,b]}|F_A(x)|^2\,dx\ \le\ \int_{[a,b]}|F_B(x)|^2\,dx∫[a,b]​∣FA​(x)∣2dx ≤ ∫[a,b]​∣FB​(x)∣2dx

holds. Here FA(x)μν=∂μAν−∂νAμ+AμAν−AνAμF_A(x)_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu+A_\mu A_\nu-A_\nu A_\muFA​(x)μν​=∂μ​Aν​−∂ν​Aμ​+Aμ​Aν​−Aν​Aμ​ at xxx, (⋆F)μν=12∑ρ,σεμνρσFρσ(\star F)_{\mu\nu}=\tfrac12\sum_{\rho,\sigma}\varepsilon_{\mu\nu\rho\sigma}F_{\rho\sigma}(⋆F)μν​=21​∑ρ,σ​εμνρσ​Fρσ​, ∣F∣2=∑μ<νRe⁡tr⁡(FμνFμν∗)|F|^2=\sum_{\mu<\nu}\operatorname{Re}\operatorname{tr}(F_{\mu\nu}F_{\mu\nu}^*)∣F∣2=∑μ<ν​Retr(Fμν​Fμν∗​), and the integrals are Lebesgue integrals over the closed box. No condition on derivatives of A,BA,BA,B at the boundary is imposed beyond equality of values there; no condition is placed on the curvature of BBB. For an empty box both sides are 000. The zero connection satisfies all hypotheses on AAA (its curvature is 000, which is both self-dual and anti-self-dual).

Human review
  • Endorsed by Shuze Chen · Oct 10, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Lucas · Oct 10, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me