Hermitian form of a spectral function in eigenbasis coordinates:
ProvedRHLinalg.hermForm_specMaplinear-algebrazeta23
Let be an Hermitian matrix over an RCLike field , with spectral decomposition ( the eigenvector unitary, the eigenvalues). For a real function , the project defines the spectral functional calculus
Statement. For every vector , writing for its eigenbasis coordinates,
That is, the Hermitian quadratic form of diagonalizes in the eigenbasis of with weights . In the module Zeta23.LinAlg.HermitianPosPart this identity drives the positivity statements about the Hermitian positive part : it is consumed by RHLinalg.posDefOn_range_hermPosPart (Sylvester, easy direction) and by Zeta23.Tail.traceNorm_le_of_hasSum_vecMulVec in the tail estimates of the matrix-variational argument.
Preamble
import Mathlib.Analysis.Matrix.PosDef
import Definitions.Def_Zeta23_LinAlg_HermitianPosPart
import Definitions.Def_Zeta23_LinAlg_PosIndex
open Matrix Finset Unitary
open scoped ComplexOrder
open RHLinalg
variable {𝕜 : Type*} [RCLike 𝕜]
variable {n : Type*} [Fintype n] [DecidableEq n]
Formal statement
theorem RHLinalg.hermForm_specMap {A : Matrix n n 𝕜} (hA : A.IsHermitian) (f : ℝ → ℝ)
(x : n → 𝕜) :
RCLike.re (star x ⬝ᵥ (specMap hA f *ᵥ x))
= ∑ i, f (hA.eigenvalues i) *
‖(star (hA.eigenvectorUnitary : Matrix n n 𝕜) *ᵥ x) i‖ ^ 2 := by sorry
Source