Lemma A.5 — (real case)
ProvedNumInvQH.Oblique.lemma_A_5linear-algebraorthogonal-projectionp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1
Let and let be linear subspaces of the Euclidean space . Write for the orthogonal complement and for the orthogonal projector onto . Then
where is the orthogonal direct sum. The lemma describes the orthogonal complement of inside as the projection of onto ; it is used with , in the proof of Theorem A.4(4).
Formalization Note The paper states the lemma in ; this is its real specialisation, on EuclideanSpace ℝ (Fin n). Each orthogonal direct sum is stated as two facts: the sum of the two subspaces is , and they are orthogonal. is Mathlib's Submodule.starProjection.
Preamble
import Mathlib
Formal statement
namespace NumInvQH.Oblique
/-- Lemma A.5, specialised from `ℂⁿ` to `ℝⁿ` (Euclidean space): for subspaces `L, M`,
`L = (L ∩ M) ⊕⊥ (L ∩ (L ∩ M)⊥) = (L ∩ M) ⊕⊥ P_L M⊥`, where `P_L` is the orthogonal projector
onto `L`. Each orthogonal direct sum is stated as "the sum equals `L` and the two summands are
orthogonal". -/
theorem lemma_A_5 {n : ℕ} (L M : Submodule ℝ (EuclideanSpace ℝ (Fin n))) :
let PL : EuclideanSpace ℝ (Fin n) →ₗ[ℝ] EuclideanSpace ℝ (Fin n) := L.starProjection
(L ⊓ M) ⊔ (L ⊓ (L ⊓ M)ᗮ) = L ∧ (L ⊓ M) ⟂ (L ⊓ (L ⊓ M)ᗮ) ∧
(L ⊓ M) ⊔ Mᗮ.map PL = L ∧ (L ⊓ M) ⟂ Mᗮ.map PL := by sorry
end NumInvQH.Oblique
Source
Černý, Czichowsky & Kallsen, Numeraire-invariant quadratic hedging and mean–variance portfolio allocation, arXiv:2110.09416v4, p. 28, Lemma A.5
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.