Algebraic polynomial interpolation of cleared derivative rows
ProvedArithmeticE.cleared_derivative_jet_interpolationdifferential-equationsformalizationlinear-algebra
Let and let be an rational polynomial matrix. Fix an algebraic complex number with . For any and any prescribed algebraic rows , there is a row of polynomials with algebraic coefficients such that
where and .
This is the finite interpolation component of the ordinary cyclic-vector construction. It involves no E-functions, convergence, or arithmetic denominator estimates. The cases and are included.
Preamble
import Definitions.Def_clearedDerivativeRows open ArithmeticE
Formal statement
theorem ArithmeticE.cleared_derivative_jet_interpolation
(m : ℕ) (T : Polynomial ℚ) (B : Matrix (Fin m) (Fin m) (Polynomial ℚ))
(ξ : ℂ) (hξ : IsAlgebraic ℚ ξ)
(hreg : T.eval₂ (algebraMap ℚ ℂ) ξ ≠ 0)
(N : ℕ) (w : ℕ → Fin m → ℂ)
(hw : ∀ k < N, ∀ i, IsAlgebraic ℚ (w k i)) :
∃ P : Fin m → Polynomial ℂ,
(∀ i k, IsAlgebraic ℚ ((P i).coeff k)) ∧
∀ k < N, ∀ i, (clearedDerivativeRows T B P k i).eval ξ = w k i := by sorrySource
Auxiliary formalization of Beukers, A refined version of the Siegel–Shidlovskii theorem, Theorem 3.2, printed pp. 6–7 (prescribed derivative rows and determinant equation), https://webspace.science.uu.nl/~beuke106/siegelshidlovskii.pdf. These explicit polynomial-numerator and module-coordinate interfaces are derived from the proof, not quoted named lemmas.