Projectors associated to Hermitian matrices #
projector: TheHermitianMatthat projects onto a given submodulesupportProj: TheHermitianMatthat projects onto the range (nonzero eigenvalues)kerProj: TheHermitianMatthat projects onto the kernelprojLE: With notation{A ≤ₚ B},projLE A Bis the projector onto the nonnegative eigenspace ofB - A.projLT: With notation{A <ₚ B},projLT A Bis the projector onto the positive eigenspace ofB - A.- Positive and negative part, written
A⁺andA⁻, give the restriction of a HermitianMat onto its positive (resp. negative) eigenvalues; equivalently, it's nonnegative (resp. nonpositive) eigenvalues.
Given a Submodule (EuclideanSpace ...) to HermitianMat, this gives the projector onto that subspace, i.e. a matrix that squares to itself, preserves vectors in the submodule, and zeroes out anything in the orthogonal complement of that submodule.
Equations
- HermitianMat.projector S = ⟨(LinearMap.toMatrix (EuclideanSpace.basisFun n 𝕜).toBasis (EuclideanSpace.basisFun n 𝕜).toBasis) ↑(S.subtypeL.comp S.orthogonalProjection), ⋯⟩
Instances For
The HermitianMat.projector for the HermitianMat.support submodule.
Equations
Instances For
The HermitianMat.projector for the HermitianMat.ker submodule.
Equations
Instances For
The projector onto a submodule S is the sum of the outer products of the vectors in an orthonormal basis of S.
The projector onto the support of A is the sum of the projections onto the eigenvectors with non-zero eigenvalues.
Projector onto the non-negative eigenspace of B - A. Accessible by the notation
{A ≤ₚ B}, which is scoped to HermitianMat. This is the unique maximum operator P
such that P^2 = P and P * A * P ≤ P * B * P in the Loewner order.
Instances For
Projector onto the positive eigenspace of B - A. Accessible by the notation
{A <ₚ B}, which is scoped to HermitianMat. Compare with proj_le.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- One or more equations did not get rendered due to their size.
Instances For
The positive part of a Hermitian matrix: the projection onto its positive eigenvalues.
Equations
- HermitianMat.instPosPart = { posPart := fun (A : HermitianMat n 𝕜) => A.cfc fun (x : ℝ) => max x 0 }
The negative part of a Hermitian matrix: the projection onto its negative eigenvalues.
Equations
- HermitianMat.instNegPart = { negPart := fun (A : HermitianMat n 𝕜) => A.cfc fun (x : ℝ) => max (-x) 0 }
There is an existing (very slow) PosPart instance on Matrix n n 𝕜, this shows
that this is equal.
There is an existing (very slow) PosPart instance on Matrix n n 𝕜, this shows
that this is equal.
The self-duality of the PSD cone: a matrix is PSD iff its inner product with all nonnegative matrices is non-negative.