Proposition 2 (i)–(ii), p. 9 — Perron–Frobenius: ρ(P) has nonnegative right and left eigenvectors, and principal blocks have smaller spectral radius
ProvedSaturatedNet.Uniqueness.proposition_2_i_iiLet be a nonnegative square matrix and let be its spectral radius (the largest modulus of a complex eigenvalue). Then:
- (if ) is an eigenvalue of , and there are vectors with
such vectors are called right and left dominant eigenvectors of ; 2. for every principal square sub-matrix of (rows and columns indexed by the same set ), .
These are the classical Perron–Frobenius facts for nonnegative matrices that the uniqueness theory rests on: (ii) gives for every class, and (i) provides the vectors , appearing in Proposition 5 and Theorem 2.
Formalization Note. Item (i) is stated under the hypothesis that the node type is nonempty (there is no nonzero vector on an empty index set; the paper's is implicit). The eigenvalue claim is carried by the nonzero eigenvector . The spectral radius is taken over in and converted to a real number for the eigen-equations; it is finite. In (ii), is allowed (its block has spectral radius ).
import Mathlib import Definitions.Def_SaturatedNet_Uniqueness_Model open Matrix
namespace SaturatedNet.Uniqueness
theorem proposition_2_i_ii {n : Type*} [Fintype n] [DecidableEq n]
(P : Matrix n n ℝ) (hP : ∀ i j, 0 ≤ P i j) :
(Nonempty n → ∃ p π : n → ℝ, 0 ≤ p ∧ p ≠ 0 ∧ 0 ≤ π ∧ π ≠ 0 ∧
P *ᵥ p = (specRad P).toReal • p ∧ π ᵥ* P = (specRad P).toReal • π) ∧
(∀ S : Finset n, specRad (block P S) ≤ specRad P) := by sorry
end SaturatedNet.Uniqueness
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.