if the have entries and has coordinates
ProvedDiazModulus.op_monomial_eval_norm_lePoints are for finite index sets . For and a weight vector on , is the operator on polynomials in the (), and for vectors () and , (the factors commute). Here the weights may be complex. Suppose every coordinate of every has absolute value at most , with ; every coordinate of has absolute value at most , with ; and . Then for , and , , where .
Proof. By induction on , for all exponents at once, peeling off the last operator. One operator sends to with . The remaining operators are linear, so the induction hypothesis bounds each term; lies below , and , make the bound increase with . Summing gives . The base case is .
Novelty. Not found in the sources read in this form; routine. An elementary size bound, used in Step 4 of the proof of DALAG Th. 11.21.
import Definitions.Def_DiazModulus
open Complex ComplexConjugate
namespace DiazModulus
theorem op_monomial_eval_norm_le {ι₀ ι₁ κ₀ : Type} [Fintype ι₀] [Fintype ι₁] [DecidableEq ι₀]
(w : κ₀ → (ι₀ → ℂ) × (ι₁ → ℂ)) (Hw : ℝ) (hHw : 0 ≤ Hw)
(hw : ∀ k, (∀ h, ‖(w k).1 h‖ ≤ Hw) ∧ ∀ i, ‖(w k).2 i‖ ≤ Hw)
(p : ι₀ → ℂ) (P : ℝ) (hP : 1 ≤ P) (hp : ∀ h, ‖p h‖ ≤ P)
(t : ι₁ → ℂ) (tb : ℝ) (ht : ∑ i, ‖t i‖ ≤ tb) (r : ℕ) (κ : Fin r → κ₀) (τ : ι₀ → ℕ) :
‖MvPolynomial.eval p ((List.ofFn fun k : Fin r =>
((∑ h, (w (κ k)).1 h • (MvPolynomial.pderiv h).toLinearMap
+ (∑ i, (w (κ k)).2 i * t i) • LinearMap.id :
Module.End ℂ (MvPolynomial ι₀ ℂ)))).prod
(∏ h, MvPolynomial.X h ^ τ h))‖
≤ (Hw * ((∑ h, τ h : ℕ) + tb + 1)) ^ r * P ^ (∑ h, τ h) := by
sorry
end DiazModulus