Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

The factorization M=P(d+a)PTM = P(d+a)P^{T}M=P(d+a)PT solves the matrix flow (51)

Proved
SDYM.matrix_flow_of_generalized_darboux_halphen

by LG · Sep 21, 2026 · Mathlib 0df444a (Lean v4.33.1)

differential-equationsintegrable-systemsmathematical-physicsmodular-forms

The reduction of the matrix flow (51) to the generalized Darboux–Halphen system, taken in the direction that does not require a diagonalizability argument. The source writes a solution of

M˙=(Adj⁡M)T+MTM−(Tr⁡M)M\dot M = (\operatorname{Adj} M)^{T} + M^{T}M - (\operatorname{Tr} M)MM˙=(AdjM)T+MTM−(TrM)M

as M=P(d+a)P−1M = P(d+a)P^{-1}M=P(d+a)P−1, where d=diag⁡(ω1,ω2,ω3)d = \operatorname{diag}(\omega_1,\omega_2,\omega_3)d=diag(ω1​,ω2​,ω3​), aaa is the skew-symmetric matrix with a12=τ3a_{12} = \tau_3a12​=τ3​, a23=τ1a_{23} = \tau_1a23​=τ1​, a31=τ2a_{31} = \tau_2a31​=τ2​, and PPP is a complex orthogonal matrix satisfying the linear equation P˙+Pa=0\dot P + Pa = 0P˙+Pa=0, equation (54). This milestone asserts that this recipe does produce a solution: if (ωj,τj)(\omega_j,\tau_j)(ωj​,τj​) solve (52)–(53) and PPP solves (54) and is complex orthogonal, then M=P(d+a)PTM = P(d+a)P^{T}M=P(d+a)PT solves (51).

Orthogonality of PPP is what makes the argument work and is not cosmetic: the right-hand side of (51) involves a transpose, and M↦(Adj⁡M)T+MTM−(Tr⁡M)MM \mapsto (\operatorname{Adj}M)^{T} + M^{T}M - (\operatorname{Tr}M)MM↦(AdjM)T+MTM−(TrM)M is equivariant under conjugation by PPP only when PT=P−1P^{T} = P^{-1}PT=P−1.

Conventions shared with the rest of the mission. All functions are complex-valued functions of a complex variable; a system is required to hold only at the points of an arbitrary set s⊆Cs \subseteq \mathbb{C}s⊆C, with no openness, connectedness or holomorphy assumed; and each derivative that a statement mentions is carried by an explicit companion function tied to it by a pointwise differentiability assertion, so no statement ever refers to the value of a derivative that does not exist.

Preamble
import Definitions.Def_SDYM_ChazyEquations
import Definitions.Def_SDYM_DarbouxHalphenRamanujan
Formal statement
namespace SDYM

theorem matrix_flow_of_generalized_darboux_halphen
    (s : Set ℂ) (w₁ w₂ w₃ x₁ x₂ x₃ : ℂ → ℂ)
    (A P M : ℂ → Matrix (Fin 3) (Fin 3) ℂ)
    (hDH : IsGeneralizedDHSolution s w₁ w₂ w₃ x₁ x₂ x₃)
    (hA : ∀ t : ℂ, A t = !![0, x₃ t, -x₂ t; -x₃ t, 0, x₁ t; x₂ t, -x₁ t, 0])
    (hPorth : ∀ t ∈ s, (P t).transpose * P t = 1)
    (hP : ∀ t ∈ s, ∀ i j, HasDerivAt (fun z => P z i j) ((-(P t * A t)) i j) t)
    (hM : ∀ t : ℂ, M t = P t * (Matrix.diagonal ![w₁ t, w₂ t, w₃ t] + A t) * (P t).transpose) :
    ∀ t ∈ s, ∀ i j, HasDerivAt (fun z => M z i j)
      (((M t).adjugate.transpose + (M t).transpose * M t
        - Matrix.trace (M t) • M t) i j) t := by sorry

end SDYM
Source
M. J. Ablowitz, S. Chakravarty, R. G. Halburd, "Integrable systems and reductions of the self-dual Yang-Mills equations", J. Math. Phys. 44 (2003) 3147-3173, https://doi.org/10.1063/1.1586967, Sec. V pp. 3164-3165, Eq. (51), the factorization M = Ms + Ma = P(d+a)P^{-1} with a12 = -a21 = τ3, a23 = -a32 = τ1, a31 = -a13 = τ2, and Eqs. (52)-(54)
Read-back

What the Lean code literally says, in plain math · Aristotle (Harmonic)

Disclosure — non-blind read-back. This read-back was written by the same agent that drafted the Lean statement it describes, with full knowledge of the source paper and of the intended meaning. It is therefore not independent testimony and must not be mistaken for an audit: the point of a read-back is that a blind reader, given only the code, renders back what the code literally says, so that any gap between intent and code becomes visible. That safeguard is absent here. A reviewer should obtain a genuinely blind read-back before relying on this item.

Fix a set s⊆Cs \subseteq \mathbb{C}s⊆C, six scalar functions w1,w2,w3,x1,x2,x3w_1,w_2,w_3,x_1,x_2,x_3w1​,w2​,w3​,x1​,x2​,x3​ on C\mathbb{C}C, and three matrix-valued functions A,P,MA, P, MA,P,M assigning to each complex ttt a 3×33\times33×3 complex matrix. There are five hypotheses.

  1. The generalized Darboux–Halphen predicate for (w1,w2,w3,x1,x2,x3)(w_1,w_2,w_3,x_1,x_2,x_3)(w1​,w2​,w3​,x1​,x2​,x3​) on sss: with T=x12+x22+x32T = x_1^2+x_2^2+x_3^2T=x12​+x22​+x32​, at every t∈st \in st∈s one has w1′=w2w3−w1(w2+w3)+Tw_1' = w_2w_3-w_1(w_2+w_3)+Tw1′​=w2​w3​−w1​(w2​+w3​)+T and its two cyclic images, and x1′=−x1(w2+w3)x_1' = -x_1(w_2+w_3)x1′​=−x1​(w2​+w3​), x2′=−x2(w3+w1)x_2' = -x_2(w_3+w_1)x2′​=−x2​(w3​+w1​), x3′=−x3(w1+w2)x_3' = -x_3(w_1+w_2)x3′​=−x3​(w1​+w2​).

  2. For every complex ttt, A(t)A(t)A(t) is the skew-symmetric matrix

A(t)=(0x3(t)−x2(t)−x3(t)0x1(t)x2(t)−x1(t)0).A(t) = \begin{pmatrix} 0 & x_3(t) & -x_2(t)\\ -x_3(t) & 0 & x_1(t) \\ x_2(t) & -x_1(t) & 0 \end{pmatrix}.A(t)=​0−x3​(t)x2​(t)​x3​(t)0−x1​(t)​−x2​(t)x1​(t)0​​.
  1. For every t∈st \in st∈s, P(t)TP(t)P(t)^{T}P(t)P(t)TP(t) is the identity matrix — that is, P(t)P(t)P(t) is complex orthogonal (with respect to the bilinear, not Hermitian, form; no conjugation is involved).

  2. For every t∈st \in st∈s and all indices i,ji,ji,j, the scalar function z↦P(z)ijz \mapsto P(z)_{ij}z↦P(z)ij​ is complex differentiable at ttt with derivative the (i,j)(i,j)(i,j) entry of − P(t)A(t)-\,P(t)A(t)−P(t)A(t).

  3. For every complex ttt, M(t)=P(t)(diag⁡(w1(t),w2(t),w3(t))+A(t))P(t)TM(t) = P(t)\bigl(\operatorname{diag}(w_1(t),w_2(t),w_3(t)) + A(t)\bigr)P(t)^{T}M(t)=P(t)(diag(w1​(t),w2​(t),w3​(t))+A(t))P(t)T.

The conclusion: for every t∈st \in st∈s and all indices i,ji,ji,j, the scalar function z↦M(z)ijz \mapsto M(z)_{ij}z↦M(z)ij​ is complex differentiable at ttt, with derivative the (i,j)(i,j)(i,j) entry of

(adj⁡M(t))T+M(t)TM(t)−(tr⁡M(t))M(t),\bigl(\operatorname{adj} M(t)\bigr)^{T} + M(t)^{T}M(t) - \bigl(\operatorname{tr} M(t)\bigr)M(t),(adjM(t))T+M(t)TM(t)−(trM(t))M(t),

where adj⁡\operatorname{adj}adj is the classical adjugate and the last term is scalar multiplication.

Observations. Hypotheses 2 and 5 are imposed at every complex ttt, whereas 1, 3 and 4 hold only on sss; this asymmetry is what makes the entrywise derivative at points of sss well defined. The conjugation in 5 uses the transpose PTP^{T}PT, not an inverse; on sss these agree by hypothesis 3, but off sss they need not. No hypothesis says P(t)P(t)P(t) is invertible off sss, and none says det⁡P(t)=1\det P(t) = 1detP(t)=1. The three off-diagonal entries of AAA are placed as A12=x3A_{12} = x_3A12​=x3​, A23=x1A_{23} = x_1A23​=x1​, A31=x2A_{31} = x_2A31​=x2​, with the opposite signs below the diagonal. If sss is empty the conclusion holds vacuously.

Human review
  • Endorsed by Shuze Chen · Sep 24, 2026

    Confirmed by the moderator at approval.

  • Endorsed by LG · Sep 24, 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