Documentation

Expdb.ExponentialSums.ExponentSumGrowthNonAsymptotic

Non-asymptotic exponential sum growth bounds #

This module formalizes the non-asymptotic definition of exponential sum exponent from the blueprint's Exponential sum growth exponents chapter (beta-chapter). It gives a fixed-parameter, epsilon--delta characterization of the exponential sum growth exponent.

Fixed-parameter formulation #

noncomputable def Expdb.exponentialSumAt (F : ) (T N : ) (a b : ) :

The exponential sum at fixed parameters.

Equations
Instances For
    def Expdb.IsApproximateModelPhaseFunction (F : ) (σ : ) (P : ) (δ : ) :

    A fixed phase function whose model-phase errors through order P are at most δ.

    Equations
    • One or more equations did not get rendered due to their size.
    Instances For

      The fixed-parameter epsilon--delta bound from the blueprint lemma beta-asymp (Non-asymptotic definition of β).

      Equations
      • One or more equations did not get rendered due to their size.
      Instances For

        Uniform approximation by fixed phases #

        From fixed bounds to asymptotic bounds #

        Uniform bounds for approximate model phases #

        theorem Expdb.approximate_model_phase_deriv_bounds {σ : } ( : 0 < σ) (P : ) :
        ∃ (K : ), 1 K ∀ {δ : } {F : }, δ min (2 ^ (-σ) / 2) 1IsApproximateModelPhaseFunction F σ P δ(∀ uphaseInterval, 2 ^ (-σ) / 2 iteratedDerivWithin 1 F phaseInterval u) ∀ (k : ), 1 kk P + 1uphaseInterval, iteratedDerivWithin k F phaseInterval u K

        An approximate model phase with sufficiently small error has a uniform positive first derivative and uniform bounds on all derivatives through order P + 1, with constants depending only on σ and P.

        The logarithm is an exact approximate model phase at every finite order.

        Building asymptotic counterexamples #

        Equivalence with the asymptotic definition #

        The non-asymptotic characterization of the exponential sum growth exponent.