§6.11, Definition consequence — Existence and uniqueness of the pseudoinverse
ProvedVectorSpaceOpt.pseudoinverse_exists_uniqueclosed-rangefunctional-analysishilbert-spaceleast-squarespseudoinverse
Let be a continuous linear map between real Hilbert spaces, and assume is closed. Then there exists a unique continuous linear map that selects, for every , the minimum-norm vector among all least-squares minimizers:
The conclusion includes more than pointwise unique minimizers: their assignment is a single bounded linear operator. This theorem validates the mission's predicate as a canonical operator construction and supplies the uniquely determined B to which the algebraic identities of Proposition 1 apply.
Preamble
import Definitions.Def_VectorSpaceOpt_pseudoinverse_predicate
Formal statement
namespace VectorSpaceOpt
theorem pseudoinverse_exists_unique
{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)) :
∃! B : H →L[ℝ] G, IsPseudoinverse A B := by sorry
end VectorSpaceOptSource
David G. Luenberger, Optimization by Vector Space Methods, §6.11, definition and existence/uniqueness discussion, book pp. 163–164 (PDF physical pp. 183–184), https://sites.science.oregonstate.edu/~show/old/142_Luenberger.pdf