Skip to content

Cayley transform

26 fully proved

Scalar operators and the Cayley transform

TheoremNonzero scalar operators are units#✓fully proved

A nonzero scalar operator is a unit.

theorem isUnit_smul_one (z : ℂ) (hz : z ≠ 0) : IsUnit (z • (1 : H →L[ℂ] H))
Used by 3

TheoremInverse of a scalar operator#✓fully proved

The ring inverse of a scalar operator is the scalar operator of the inverse scalar.

theorem ringInverse_smul_one (z : ℂ) :
    Ring.inverse (z • (1 : H →L[ℂ] H)) = z⁻¹ • (1 : H →L[ℂ] H)
Uses 1

DefinitionCayley transform of an operator#✓fully proved

The Cayley transform of an operator A: (A - i) * (A + i)⁻¹. The factor A + i • 1 is a unit when A is self-adjoint on a complete Hilbert space (see isUnit_add_I_smul_one_of_isSelfAdjoint below), in which case the ring inverse here is the genuine operator inverse.

noncomputable def cayleyTransform : H →L[ℂ] H :=
  (A - Complex.I • (1 : H →L[ℂ] H)) * Ring.inverse (A + Complex.I • (1 : H →L[ℂ] H))
Used by 18

TheoremFactors of the Cayley transform commute#✓fully proved

The two factors of the Cayley transform commute.

theorem commute_sub_I_add_I :
    (A - Complex.I • (1 : H →L[ℂ] H)) * (A + Complex.I • (1 : H →L[ℂ] H))
      = (A + Complex.I • (1 : H →L[ℂ] H)) * (A - Complex.I • (1 : H →L[ℂ] H))
Used by 2

Self Adjoint

TheoremSelf-adjoint operator minus a non-real scalar is invertible#✓fully proved

For a self-adjoint operator, A - w • 1 is a unit whenever w has nonzero imaginary part. The proof is the norm estimate ‖(A - w • 1) x‖ ≥ |w.im| * ‖x‖, read off from the imaginary part of ⟪(A - w • 1) x, x⟫.

theorem isUnit_sub_smul_one_of_isSelfAdjoint_of_im_ne_zero (hA : IsSelfAdjoint A)
    {w : ℂ} (hw : w.im ≠ 0) : IsUnit (A - w • (1 : H →L[ℂ] H))
Used by 2

TheoremSelf-adjoint operator plus \(i\) is invertible#✓fully proved

For a self-adjoint operator, A + i • 1 is a unit.

theorem isUnit_add_I_smul_one_of_isSelfAdjoint (hA : IsSelfAdjoint A) :
    IsUnit (A + Complex.I • (1 : H →L[ℂ] H))
Uses 1 · Used by 8

TheoremSelf-adjoint operator minus \(i\) is invertible#✓fully proved

For a self-adjoint operator, A - i • 1 is a unit.

theorem isUnit_sub_I_smul_one_of_isSelfAdjoint (hA : IsSelfAdjoint A) :
    IsUnit (A - Complex.I • (1 : H →L[ℂ] H))
Uses 1 · Used by 3

TheoremInverse of a self-adjoint operator plus \(i\) is a unit#✓fully proved

The inverse of A + i • 1 is a unit for self-adjoint A.

theorem isUnit_ringInverse_add_I_of_isSelfAdjoint (hA : IsSelfAdjoint A) :
    IsUnit (Ring.inverse (A + Complex.I • (1 : H →L[ℂ] H)))
Uses 1 · Used by 4

TheoremAdjoint of the Cayley transform#✓fully proved

The adjoint of the Cayley transform.

theorem star_cayleyTransform (hA : IsSelfAdjoint A) :
    star (cayleyTransform A)
      = Ring.inverse (A - Complex.I • (1 : H →L[ℂ] H))
          * (A + Complex.I • (1 : H →L[ℂ] H))
Uses 1 · Used by 2

TheoremAdjoint of the Cayley transform is a left inverse#✓fully proved

The Cayley transform of a self-adjoint operator satisfies U† * U = 1.

theorem star_mul_self_cayleyTransform (hA : IsSelfAdjoint A) :
    star (cayleyTransform A) * cayleyTransform A = 1
Uses 5 · Used by 1

TheoremAdjoint of the Cayley transform is a right inverse#✓fully proved

The Cayley transform of a self-adjoint operator satisfies U * U† = 1.

theorem mul_star_self_cayleyTransform (hA : IsSelfAdjoint A) :
    cayleyTransform A * star (cayleyTransform A) = 1
Uses 5

TheoremCayley transform of a self-adjoint operator is invertible#✓fully proved

The Cayley transform of a self-adjoint operator is a unit.

theorem isUnit_cayleyTransform (hA : IsSelfAdjoint A) : IsUnit (cayleyTransform A)
Uses 3 · Used by 1

TheoremCayley transform of a self-adjoint operator is unitary#✓fully proved

The Cayley transform of a self-adjoint operator is unitary.

theorem mem_unitary_cayleyTransform (hA : IsSelfAdjoint A) :
    cayleyTransform A ∈ unitary (H →L[ℂ] H)
Uses 3 · Used by 2

TheoremCayley transform preserves norms of vectors#✓fully proved

The Cayley transform of a self-adjoint operator is an isometry.

theorem norm_cayleyTransform_apply (hA : IsSelfAdjoint A) (x : H) :
    ‖cayleyTransform A x‖ = ‖x‖
Uses 2 · Used by 1

TheoremOperator norm of the Cayley transform is one#✓fully proved

On a nontrivial space, the Cayley transform has operator norm 1.

theorem norm_cayleyTransform (hA : IsSelfAdjoint A) [Nontrivial H] :
    ‖cayleyTransform A‖ = 1
Uses 2

TheoremFormula for one plus the Cayley transform#✓fully proved

1 + U is 2 • A times the inverse of A + i • 1.

theorem one_add_cayleyTransform (hA : IsSelfAdjoint A) :
    1 + cayleyTransform A
      = (2 : ℂ) • A * Ring.inverse (A + Complex.I • (1 : H →L[ℂ] H))
Uses 2 · Used by 1

TheoremFormula for one minus the Cayley transform#✓fully proved

1 - U is 2 • (i • 1) times the inverse of A + i • 1.

theorem one_sub_cayleyTransform (hA : IsSelfAdjoint A) :
    1 - cayleyTransform A
      = (2 : ℂ) • (Complex.I • (1 : H →L[ℂ] H))
          * Ring.inverse (A + Complex.I • (1 : H →L[ℂ] H))
Uses 2 · Used by 2

TheoremOne minus the Cayley transform is invertible#✓fully proved

1 - U is a unit for the Cayley transform of a self-adjoint operator: 1 is never in the spectrum of U.

theorem isUnit_one_sub_cayleyTransform (hA : IsSelfAdjoint A) :
    IsUnit (1 - cayleyTransform A)
Uses 3 · Used by 2

TheoremInverse Cayley formula#✓fully proved

The inverse Cayley formula: A = i • (1 + U) * (1 - U)⁻¹.

theorem eq_cayleyTransform_inv (hA : IsSelfAdjoint A) :
    A = Complex.I • (1 + cayleyTransform A) * Ring.inverse (1 - cayleyTransform A)
Uses 4

TheoremResolvent factorization for the Cayley transform#✓fully proved

The resolvent-type factorization behind the spectral mapping: for μ ≠ 1 and λ = i * (1 + μ) / (1 - μ), μ • 1 - U = ((μ - 1) • (A - λ • 1)) * (A + i • 1)⁻¹.

theorem algebraMap_sub_cayleyTransform (hA : IsSelfAdjoint A) {μ : ℂ} (hμ : μ ≠ 1) :
    algebraMap ℂ (H →L[ℂ] H) μ - cayleyTransform A
      = ((μ - 1) • (A - (Complex.I * (1 + μ) / (1 - μ)) • (1 : H →L[ℂ] H)))
          * Ring.inverse (A + Complex.I • (1 : H →L[ℂ] H))
Uses 2 · Used by 2

TheoremOne is not in the spectrum of the Cayley transform#✓fully proved

1 is never in the spectrum of the Cayley transform.

theorem one_notMem_spectrum_cayleyTransform (hA : IsSelfAdjoint A) :
    (1 : ℂ) ∉ spectrum ℂ (cayleyTransform A)
Uses 2 · Used by 1

TheoremSpectral mapping for the Cayley transform, image direction#✓fully proved

The image direction of the spectral mapping: if w is in the spectrum of A, then (w - i) / (w + i) is in the spectrum of U.

theorem mem_spectrum_cayleyTransform_of_mem_spectrum (hA : IsSelfAdjoint A)
    {w : ℂ} (hw : w ∈ spectrum ℂ A) :
    (w - Complex.I) / (w + Complex.I) ∈ spectrum ℂ (cayleyTransform A)
Uses 5 · Used by 1

TheoremSpectral mapping for the Cayley transform, preimage direction#✓fully proved

The preimage direction of the spectral mapping: every point of the spectrum of U is (w - i) / (w + i) for some w in the spectrum of A.

theorem exists_mem_spectrum_of_mem_spectrum_cayleyTransform (hA : IsSelfAdjoint A)
    {μ : ℂ} (hμ : μ ∈ spectrum ℂ (cayleyTransform A)) :
    ∃ w ∈ spectrum ℂ A, (w - Complex.I) / (w + Complex.I) = μ
Uses 5 · Used by 1

TheoremSpectrum of the Cayley transform#✓fully proved

The spectrum of the Cayley transform is the image of the spectrum of A under the Möbius map λ ↦ (λ - i) / (λ + i).

theorem spectrum_cayleyTransform (hA : IsSelfAdjoint A) :
    spectrum ℂ (cayleyTransform A)
      = (fun w => (w - Complex.I) / (w + Complex.I)) '' spectrum ℂ A
Uses 3

TheoremSpectrum of the Cayley transform lies on the unit circle#✓fully proved

Points of the spectrum of the Cayley transform lie on the unit circle.

theorem norm_eq_one_of_mem_spectrum_cayleyTransform (hA : IsSelfAdjoint A)
    {μ : ℂ} (hμ : μ ∈ spectrum ℂ (cayleyTransform A)) : ‖μ‖ = 1
Uses 2

TheoremEigenvalues of the Cayley transform#✓fully proved

The eigenvalue correspondence: for μ ≠ 1, μ is an eigenvalue of the Cayley transform U iff i * (1 + μ) / (1 - μ) is an eigenvalue of A.

theorem hasEigenvalue_cayleyTransform_iff (hA : IsSelfAdjoint A) {μ : ℂ} (hμ : μ ≠ 1) :
    Module.End.HasEigenvalue ((cayleyTransform A : H →L[ℂ] H) : Module.End ℂ H) μ ↔
      Module.End.HasEigenvalue ((A : H →L[ℂ] H) : Module.End ℂ H)
        (Complex.I * (1 + μ) / (1 - μ))
Uses 2

The dependency explorer needs JavaScript and its graph data.

Download the complete graph data.