Square-summable conjugate-symmetric sequences are Fourier coefficients of functions
ProvedAsymptoticOperator.exists_coeff_eqLet with its Haar probability measure and let . For and let be its -th Fourier coefficient, computed componentwise after viewing the real entries of as complex numbers.
Let satisfy and (componentwise) for every . Then there is with for every .
Together with the uniqueness of Fourier coefficients, this identifies with the conjugate-symmetric square-summable sequences in . It is used to produce elements of from prescribed Fourier coefficients, for example the image of an element of .
Formalization note: L2 n and coeff come from the definition module AsymptoticOperator_Setting; L2 n is Lp (EuclideanSpace ℝ (Fin n ⊕ Fin n)) 2 over UnitAddCircle with AddCircle.haarAddCircle, and coeff h k is the vector . Square-summability is stated as summability of , and conjugate symmetry componentwise with starRingEnd ℂ.
import Definitions.Def_AsymptoticOperator_Setting
namespace AsymptoticOperator
open ConleyZehnder
/-- Every square-summable, conjugate-symmetric family `c : ℤ → ℂ²ⁿ` is the sequence of
Fourier coefficients of an element of `L²(S¹, ℝ²ⁿ)`. -/
theorem exists_coeff_eq {n : ℕ} (c : ℤ → Fin n ⊕ Fin n → ℂ)
(hsum : Summable fun k : ℤ => ∑ i, ‖c k i‖ ^ 2)
(hconj : ∀ k i, c (-k) i = (starRingEnd ℂ) (c k i)) :
∃ h : L2 n, ∀ k : ℤ, coeff h k = c k := by sorry
end AsymptoticOperator