§4.1 — along the gradient flow,
ProvedFatkhullinPolyak.Flow.flow_energy_identitygradient-flowlqrp2o-batch-p100bp2o-gran-per-chapterp2o-plan-paperp2o-v1
Let and let be a solution on of the gradient flow (4.1), , (in particular for all ). Then for every the function is differentiable at (from the right at ) with
This energy identity is "the main idea of the proof" of Theorem 4.1: it shows at once that is nonincreasing.
Formalization Note The derivative is taken within . The norm is the Frobenius norm, the one with respect to which (3.3) is the gradient.
Preamble
import Mathlib import Definitions.Def_FatkhullinPolyak_Flow_LQR import Definitions.Def_FatkhullinPolyak_Flow_GradientFlow
Formal statement
namespace FatkhullinPolyak.Flow
open Matrix
theorem flow_energy_identity {n m r : ℕ}
(A : Matrix (Fin n) (Fin n) ℝ) (B : Matrix (Fin n) (Fin m) ℝ) (C : Matrix (Fin r) (Fin n) ℝ)
(Q : Matrix (Fin n) (Fin n) ℝ) (R : Matrix (Fin m) (Fin m) ℝ) (Sig : Matrix (Fin n) (Fin n) ℝ)
(hQ : Q.PosDef) (hR : R.PosDef) (hSig : Sig.PosDef)
(K₀ : Matrix (Fin m) (Fin r) ℝ) (K : ℝ → Matrix (Fin m) (Fin r) ℝ)
(hK : IsGradFlow A B C Q R Sig K₀ K) :
∀ t : ℝ, 0 ≤ t →
HasDerivWithinAt (fun s => cost A B C Q R Sig (K s))
(-(frobNorm (grad A B C Q R Sig (K t)) ^ 2)) (Set.Ici 0) t := by sorry
end FatkhullinPolyak.Flow
Source
Fatkhullin, Polyak, Optimizing Static Linear Feedback: Gradient Method, arXiv:2004.09875v2, p. 10, §4.1, last paragraph; Appendix D.1, p. 18
Read-back
What the Lean code literally says, in plain math · claude-opus-5-5
Setting. Let
Gains are real matrices. The notions used are:
- is the set of gains with every complex eigenvalue of having negative real part.
- and are the unique solutions (or if the solution is not unique) of
- The cost is .
- is the explicit formula.
- is the Frobenius norm.
Hypotheses.
- , , are positive definite (symmetric, with for ).
- , and is a gradient-flow solution from . This means:
- ;
- for all ;
- for every , each entry has derivative at relative to (one-sided at ).
- No rank condition on and no condition on is imposed.
Conclusion. For every , the function has derivative at relative to equal to
At this is a right derivative.
Degenerate cases.
- If no gradient-flow solution exists for the data, the statement is vacuous.
- If , then and , and the claim is that a constant function has derivative .
- If or , the curve is constant on and is the empty matrix with norm .
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.