Proof of Proposition 4.16, p. 17 —
ProvedQuadMatIneq.Petersen.set_identitynorm-bounded-uncertaintyp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1petersen-lemmaquadratic-matrix-inequality
Let with , let be its positive semidefinite square root, and let . Then
and this set equals
This is the first step of the paper's proof of Petersen's lemma: it replaces the products over the norm-bounded uncertainty set by the solution set of a single quadratic matrix inequality in .
Formalization Note The square root is Mathlib's CFC.sqrt for the (scoped) matrix order, which on a positive semidefinite matrix is its unique positive semidefinite square root. The sets are written as images of and .
Preamble
import Mathlib import Definitions.Def_QuadMatIneq_Petersen_QMI import Definitions.Def_QuadMatIneq_Petersen_Setup open scoped MatrixOrder
Formal statement
namespace QuadMatIneq.Petersen
open Matrix
theorem set_identity {n p q : ℕ} (Fbar : Matrix (Fin q) (Fin q) ℝ) (G : Matrix (Fin q) (Fin n) ℝ)
(hFbar : Fbar.PosSemidef) :
(fun F : Matrix (Fin p) (Fin q) ℝ => F * G) '' uncSet Fbar =
(fun S : Matrix (Fin p) (Fin q) ℝ => S * CFC.sqrt Fbar * G) ''
{S | (1 - Sᵀ * S).PosSemidef} ∧
(fun F : Matrix (Fin p) (Fin q) ℝ => F * G) '' uncSet Fbar =
{H : Matrix (Fin p) (Fin n) ℝ | (Gᵀ * Fbar * G - Hᵀ * H).PosSemidef} := by sorry
end QuadMatIneq.Petersen
Source
van Waarde–Camlibel–Eising–Trentelman, arXiv:2203.12959v3, §4.5.3, proof of Proposition 4.16, p. 17, first paragraph (the two set identities obtained from Lemma A.1)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.