Polynomial commutant bounds for monomial word actions #
Matrix entries of an intertwiner are determined, up to nonzero
scalars, by the simultaneous permutation orbits of pairs of words.
Such an orbit is determined by its letter-pair counts. Each count
lies between zero and the word length, giving the polynomial bound
(n + 1) ^ (Fintype.card α ^ 2).
structure
RS.MonomialWordAction
{α : Type u_1}
{n : ℕ}
(ρ : Representation ℂ (Equiv.Perm (Fin n)) ((Fin n → α) → ℂ))
:
Type u_1
Coordinates of a symmetric-group representation acting on words by permutation of positions and multiplication by nonzero scalars.
- weight : Equiv.Perm (Fin n) → (Fin n → α) → ℂ
The scalar attached to a permutation and an output word.
Every coordinate scalar is nonzero.
- apply_eq (σ : Equiv.Perm (Fin n)) (v : (Fin n → α) → ℂ) (c : Fin n → α) : (ρ σ) v c = self.weight σ c * v (c ∘ ⇑σ)
The action reindexes coordinates by the permutation.
Instances For
theorem
RS.finrank_commutant_le_word_counts
{α : Type u_1}
{n : ℕ}
[Fintype α]
{ρ : Representation ℂ (Equiv.Perm (Fin n)) ((Fin n → α) → ℂ)}
(M : MonomialWordAction ρ)
:
The commutant of a monomial word action has polynomial dimension, with one possible coordinate for each table of letter-pair counts.