AFTD Auto-Formalizing Theoretical Domains

laplacian_order_four_stationary_eigenvalue_sector

theoremverified

For a real order-four matrix with zero row sums, nonpositive off-diagonal entries, and a strictly positive left null vector, every complex eigenvalue satisfies the sector inequality.

Statement

theorem laplacian_order_four_stationary_eigenvalue_sector
    (L : Matrix (Fin 4) (Fin 4) ℝ) (π : Fin 4 → ℝ)
    (hπ : ∀ i, 0 < π i)
    (hr : ∀ i, ∑ j, L i j = 0)
    (ho : ∀ i j, i ≠ j → L i j ≤ 0)
    (hs : ∀ j, ∑ i, π i * L i j = 0)
    (v : Fin 4 → ℂ) (hv0 : v ≠ 0) (z : ℂ)
    (hv : (L.map (algebraMap ℝ ℂ)).mulVec v = z • v) :
    |z.im| ≤ z.re
Source
Sector-bound formalization for IN-166; arXiv:math/0508176, Theorem 5.
Verified
27 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 1
Used by 1

The proof, explained

The proof establishes the sector inequality |Im(z)| ≤ Re(z) by combining two one-sided bounds. The helper theorem laplacian_order_four_stationary_eigenvalue_one_side provides the first bound, 0 ≤ Re(z) + Im(z). Because L is a real matrix, the complex conjugate vector v̄ is an eigenvector with eigenvalue z̄; applying the one-sided theorem to this conjugate pair yields 0 ≤ Re(z) - Im(z). Together, these two linear inequalities bound Im(z) between -Re(z) and Re(z), proving |Im(z)| ≤ Re(z).

  1. The eigenvalue z satisfies -Re(z) ≤ Im(z), or equivalently 0 ≤ Re(z) + Im(z).

    The helper theorem laplacian_order_four_stationary_eigenvalue_one_side establishes that every eigenvalue z of an order-4 Laplacian-like matrix with a strictly positive left null vector satisfies 0 ≤ Re(z) + Im(z). Applying it directly to the eigenvector v and eigenvalue z yields this lower bound.

    line 10

    uses laplacian_order_four_stationary_eigenvalue_one_side

      have hp := laplacian_order_four_stationary_eigenvalue_one_side L π hπ hr ho hs v hv0 z hv
  2. The entrywise complex conjugate vector v̄ is an eigenvector of L with eigenvalue z̄.

    Extending equality coordinate-wise (ext i), we take the complex conjugate of both sides of the i-th coordinate equation (congrFun and congrArg). The simpa tactic distributes complex conjugation across the dot product and sum defining matrix-vector multiplication (Matrix.mulVec, dotProduct, map_sum, map_mul). Since the entries of L are real, they are fixed by conjugation, proving that L v̄ = z̄ v̄.

    lines 11–14

    uses congrArg, congrFun, Matrix.mulVec, dotProduct, map_sum, map_mul

      have hvc : (L.map (algebraMap ℝ ℂ)).mulVec (star v) = (star z) • (star v) := by
        ext i
        have h := congrArg star (congrFun hv i)
        simpa [Matrix.mulVec, dotProduct, map_sum, map_mul] using h
  3. The eigenvalue z satisfies Im(z) ≤ Re(z), or equivalently 0 ≤ Re(z) - Im(z).

    Since v ≠ 0, its conjugate v̄ is also nonzero. We apply laplacian_order_four_stationary_eigenvalue_one_side to the conjugate eigenpair (v̄, z̄) to obtain 0 ≤ Re(z̄) + Im(z̄). Rewriting with Complex.star_def, Complex.conj_re, and Complex.conj_im transforms Re(z̄) into Re(z) and Im(z̄) into -Im(z), yielding 0 ≤ Re(z) - Im(z), which is Im(z) ≤ Re(z).

    lines 15–17

    uses laplacian_order_four_stationary_eigenvalue_one_side, Complex.star_def, Complex.conj_re, Complex.conj_im

      have hm := laplacian_order_four_stationary_eigenvalue_one_side L π hπ hr ho hs (star v)
        (by simpa using hv0) (star z) hvc
      simp only [Complex.star_def, Complex.conj_re, Complex.conj_im] at hm
  4. The eigenvalue z satisfies the sector inequality |Im(z)| ≤ Re(z).

    By abs_le.mpr, |Im(z)| ≤ Re(z) is equivalent to the two inequalities -Re(z) ≤ Im(z) and Im(z) ≤ Re(z). These were established in the preceding steps, and linear arithmetic (linarith) verifies each branch.

    line 18

    uses abs_le.mpr

      exact abs_le.mpr ⟨by linarith, by linarith⟩
laplacian_order_four_stationary_eigenvalue_one_side
for a real 4×4 matrix with zero row sums, nonpositive off-diagonal entries, and a strictly positive stationary vector, any complex eigenvalue z satisfies 0 ≤ Re(z) + Im(z)
congrArg
if two arguments are equal, applying the same function to both yields equal results: a₁ = a₂ → f(a₁) = f(a₂)
congrFun
if two functions are equal, their evaluations at any point are equal: f = g → f(a) = g(a)
Matrix.mulVec
matrix-vector multiplication, defined by taking the dot product of each matrix row with the vector
dotProduct
the standard dot product of two vectors, defined as the sum of their coordinate-wise products
map_sum
an additive monoid homomorphism preserves finite sums: g(∑ x ∈ s, f x) = ∑ x ∈ s, g(f x)
map_mul
a multiplicative map preserves products: f(x * y) = f(x) * f(y)
Complex.star_def
the star operation on ℂ is given by complex conjugation (starRingEnd ℂ)
Complex.conj_re
the real part of the complex conjugate of z equals the real part of z: Re(z̄) = Re(z)
Complex.conj_im
the imaginary part of the complex conjugate of z is the negation of the imaginary part of z: Im(z̄) = -Im(z)
abs_le.mpr
if -b ≤ a and a ≤ b, then |a| ≤ b

Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.

Lean source view module on GitHub

theorem laplacian_order_four_stationary_eigenvalue_sector    (L : Matrix (Fin 4) (Fin 4) ℝ) (π : Fin 4 → ℝ)    (hπ : ∀ i, 0 < π i)    (hr : ∀ i, ∑ j, L i j = 0)    (ho : ∀ i j, i ≠ j → L i j ≤ 0)    (hs : ∀ j, ∑ i, π i * L i j = 0)    (v : Fin 4 → ℂ) (hv0 : v ≠ 0) (z : ℂ)    (hv : (L.map (algebraMap ℝ ℂ)).mulVec v = z • v) :    |z.im| ≤ z.re := by  have hp := laplacian_order_four_stationary_eigenvalue_one_side L π hπ hr ho hs v hv0 z hv  have hvc : (L.map (algebraMap ℝ ℂ)).mulVec (star v) = (star z) • (star v) := by    ext i    have h := congrArg star (congrFun hv i)    simpa [Matrix.mulVec, dotProduct, map_sum, map_mul] using h  have hm := laplacian_order_four_stationary_eigenvalue_one_side L π hπ hr ho hs (star v)    (by simpa using hv0) (star z) hvc  simp only [Complex.star_def, Complex.conj_re, Complex.conj_im] at hm  exact abs_le.mpr ⟨by linarith, by linarith⟩
See it in the graphDiscuss this resultChallenge it