§6.6, Theorem 2 — Closed-range adjoint identity
ProvedVectorSpaceOpt.closed_range_adjoint_range_eq_orthogonal_kerLet and be real Hilbert spaces, let be continuous linear, and assume is closed. Then the range of the Hilbert adjoint is exactly the orthogonal complement of the kernel of :
The closedness hypothesis is essential for equality rather than only a closure relation. This theorem identifies the canonical minimum-norm component of a solution as an adjoint-range vector and is the range/kernel bridge used by the minimum-norm characterization and the construction of a bounded pseudoinverse throughout Hilbert-space optimization.
import Mathlib open scoped InnerProduct
namespace VectorSpaceOpt
theorem closed_range_adjoint_range_eq_orthogonal_ker
{G H : Type*} [NormedAddCommGroup G] [InnerProductSpace ℝ G] [CompleteSpace G]
[NormedAddCommGroup H] [InnerProductSpace ℝ H] [CompleteSpace H]
(A : G →L[ℝ] H) (hA : IsClosed (A.range : Set H)) :
(A†).range = A.kerᗮ := by sorry
end VectorSpaceOptRead-back
What the Lean code literally says, in plain math · gpt-5
Let and be complete real inner-product spaces, and let be a continuous real-linear map whose range, as a subset of , is closed. Then the range of the adjoint , as a linear subspace of , is exactly the orthogonal complement in of the kernel of . This is literal subspace equality, with no additional closure placed around the range of , and no finite-dimensionality or nontriviality assumption is made.
Confirmed by the mission captain (proposal self-audit).