Weyl's inequality: sorted eigenvalues of symmetric operators are monotone in the form order
ProvedLinearMap.IsSymmetric.eigenvalues_le_add_of_re_inner_leeigenvalueslinear-algebraspectral-theory
Let be an inner product space over of finite dimension , , and symmetric linear maps with
Then, with eigenvalues sorted in decreasing order, for every .
This is the operator form of Weyl's monotonicity inequality; the matrix form for Hermitian matrices is published separately.
Formalization Note: hT.eigenvalues hn : Fin m → ℝ is Mathlib's LinearMap.IsSymmetric.eigenvalues, which is antitone (decreasing order).
Preamble
import Mathlib
Formal statement
namespace LinearMap.IsSymmetric
open Module
/-- Let `T` and `S` be symmetric operators on an `m`-dimensional inner product space over `ℝ` or
`ℂ`, with `Re ⟪T x, x⟫ ≤ Re ⟪S x, x⟫ + ε ‖x‖²` for all `x`. Then for each `j`, the `j`-th
eigenvalue of `T` (eigenvalues in decreasing order) is at most the `j`-th eigenvalue of `S`
plus `ε`. -/
theorem eigenvalues_le_add_of_re_inner_le {𝕜 : Type*} [RCLike 𝕜] {E : Type*}
[NormedAddCommGroup E] [InnerProductSpace 𝕜 E] [FiniteDimensional 𝕜 E] {T S : E →ₗ[𝕜] E}
{m : ℕ} (hT : T.IsSymmetric) (hS : S.IsSymmetric) (hn : finrank 𝕜 E = m) (ε : ℝ)
(h : ∀ x : E, RCLike.re (inner 𝕜 (T x) x) ≤ RCLike.re (inner 𝕜 (S x) x) + ε * ‖x‖ ^ 2)
(j : Fin m) :
hT.eigenvalues hn j ≤ hS.eigenvalues hn j + ε := by
sorry
end LinearMap.IsSymmetric
Source
Weyl (1912); standard