Documentation

LeanPool.MatrixMultiplicationAllFields.LinearAlgebra.MatrixMultiplication.AuxiliarySeparation.Spectrum.MultiplicativeStates

Multiplicative states obtained by normalizing translates of additive states.

Translate a state by multiplication, then normalize at the unit.

Equations
Instances For

    Functions satisfying a fixed multiplicative scaling relation at one semiring element.

    Equations
    Instances For
      theorem MatrixMultiplicationAllFields.MatrixMultiplication.AuxiliarySeparation.normalizedTranslate_eq_mem_frozenSlice {S : Type u_1} [CommSemiring S] {C : Set (S → ℝ)} {z : S} {f : S → ℝ} (hfC : f ∈ C) (hf1 : f 1 = 1) (hf : normalizedTranslate z f = f) :
      f ∈ frozenSlice C z (f z)
      theorem MatrixMultiplicationAllFields.MatrixMultiplication.AuxiliarySeparation.exists_finite_multiplicative_slice {S : Type u_1} [CommSemiring S] {K : Set (S → ℝ)} (hKcompact : IsCompact K) (hKconvex : Convex ℝ K) (hKne : K.Nonempty) (hKnorm : ∀ f ∈ K, f 1 = 1) (hKnonzero : ∀ f ∈ K, ∀ (z : S), z ≠ 0 → f z ≠ 0) (hKmap : ∀ (z : S), z ≠ 0 → Set.MapsTo (normalizedTranslate z) K K) (t : Finset { z : S // z ≠ 0 }) :
      ∃ C ⊆ K, C.Nonempty ∧ IsCompact C ∧ Convex ℝ C ∧ (∀ (z : S), z ≠ 0 → Set.MapsTo (normalizedTranslate z) C C) ∧ ∀ f ∈ C, ∀ z ∈ t, ∀ (x : S), f (↑z * x) = f ↑z * f x

      Every finite collection of multiplication identities holds throughout a nonempty compact convex subset that remains invariant under all translates.

      theorem MatrixMultiplicationAllFields.MatrixMultiplication.AuxiliarySeparation.exists_multiplicative_of_invariant_states {S : Type u_1} [CommSemiring S] {K : Set (S → ℝ)} (hKcompact : IsCompact K) (hKconvex : Convex ℝ K) (hKne : K.Nonempty) (hKnorm : ∀ f ∈ K, f 1 = 1) (hKzero : ∀ f ∈ K, f 0 = 0) (hKnonzero : ∀ f ∈ K, ∀ (z : S), z ≠ 0 → f z ≠ 0) (hKmap : ∀ (z : S), z ≠ 0 → Set.MapsTo (normalizedTranslate z) K K) :
      ∃ f ∈ K, ∀ (x y : S), f (x * y) = f x * f y

      A compact convex normalized state space invariant under multiplication translates contains a multiplicative state. Frozen eigenvalue slices preserve convexity at each finite fixed-point step.

      theorem MatrixMultiplicationAllFields.MatrixMultiplication.AuxiliarySeparation.normalizedStates_exists_multiplicative {S : Type u_1} [CommSemiring S] [Preorder S] (R : S → ℕ) (d : S) (k : ℝ) (hunit : ∀ (z : S), z ≠ 0 → 1 ≤ z) (hmulmono : ∀ (z x y : S), x ≤ y → z * x ≤ z * y) (hdom : ∀ (z x : S), z * x ≤ ↑(R x) * z) (hne : (normalizedStates (fun (x : S) => ↑(R x)) d k).Nonempty) :
      ∃ f ∈ normalizedStates (fun (x : S) => ↑(R x)) d k, ∀ (x y : S), f (x * y) = f x * f y

      Appendix A.2 for a ranked restriction semiring: once the preliminary normalized state set is nonempty, it contains a multiplicative state.

      theorem MatrixMultiplicationAllFields.MatrixMultiplication.AuxiliarySeparation.exists_monotone_semiringHom_of_normalizedStates {S : Type u_1} [CommSemiring S] [Preorder S] (R : S → ℕ) (d : S) (k : ℝ) (hunit : ∀ (z : S), z ≠ 0 → 1 ≤ z) (hmulmono : ∀ (z x y : S), x ≤ y → z * x ≤ z * y) (hdom : ∀ (z x : S), z * x ≤ ↑(R x) * z) (hne : (normalizedStates (fun (x : S) => ↑(R x)) d k).Nonempty) :
      ∃ (χ : S →+* ℝ), Monotone ⇑χ ∧ (∀ (x : S), 0 ≤ χ x ∧ χ x ≤ ↑(R x)) ∧ k ≤ χ d

      The multiplicative state as a normalized monotone semiring homomorphism, with the original rank bounds and detector inequality preserved.