Uniform seminorm bounds for families of exact Fourier aliases #
All constants are chosen before the source and the band. The estimates use actual derivatives and actual translated integrals.
Actual torus inversion for smooth finite-dimensional parameter families #
The source is an actual jointly smooth function. Fourier coefficients are the unit-square integrals of that source. No output regularity or decay assumptions are part of the construction.
Point: an abbreviation for P × Plane.
Instances For
Slice, defined pointwise by f (p, Y).
Equations
- NavierStokes.SmoothFamilyTorusInverse.slice f p Y = f (p, Y)
Instances For
Periodic, given by ∀ p, SmoothFourierData.UnitPeriodic (slice f p).
Equations
Instances For
Coefficient, given by SmoothFourierData.coefficient (slice f p) k.
Equations
Instances For
Inverse, given by directionalInverse d (coefficient f z.1) z.2.
Equations
Instances For
Fixed partial, given by fderiv ℝ f z v.
Equations
- NavierStokes.SmoothFamilyTorusInverse.fixedPartial v f z = (fderiv ℝ f z) v
Instances For
Parameter partial, given by fixedPartial (v, 0) f.
Equations
Instances For
Parameter derivative, given by (fderiv ℝ f z).comp (ContinuousLinearMap.inl ℝ P Plane).
Equations
Instances For
Torus X jet, given by (fixedPartial (0, (1, 0)))^[n] f.
Equations
Instances For
Weighted source, given by SmoothFourierData.kernel k (q.2, q.1.2) * f (q.1.1, (q.2, q.1.2)).
Equations
Instances For
Restricting the actual family to a line only changes its external parameter. It leaves every torus Fourier integral unchanged.
Instances For
Apply multiplier, given by series (fun k => m k * coefficient f z.1 k) z.2.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Repeat partial, given by (fun g x => fderiv ℝ g x v)^[n] f.
Equations
Instances For
Tensor torus word as an element of w, g => fun Y => fderiv ℝ (tensorTorusWord w g) Y (if b then (0, 1) else (1, 0)).
Equations
- One or more equations did not get rendered due to their size.
- NavierStokes.SmoothFamilyTorusInverse.tensorTorusWord [] x✝ = x✝
Instances For
Basis index: an abbreviation for Fin (Module.finrank ℝ P).
Equations
Instances For
Parameter basis, given by Module.finBasis ℝ P.
Instances For
Parameter coord, given by ((parameterBasis (P := P)).coord i).toContinuousLinearMap.
Equations
Instances For
Parameter lift, given by ContinuousLinearMap.smulRightL ℝ (Point P) ℂ ((parameterCoord i).comp (ContinuousLinearMap.fst ℝ P Plane)).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Torus lift X, given by ContinuousLinearMap.smulRightL ℝ (Point P) ℂ (dx.comp (ContinuousLinearMap.snd ℝ P Plane)).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Torus lift Y, given by ContinuousLinearMap.smulRightL ℝ (Point P) ℂ (dy.comp (ContinuousLinearMap.snd ℝ P Plane)).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Multiplier term derivative as an element of Point P →L[ℝ] ℂ.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Directional partial, given by fderiv ℝ f z (0, vector d).
Equations
Instances For
A prefix of actual full joint Fréchet-jet bounds on a parameter set. The torus argument ranges over the whole universal cover.
Equations
- NavierStokes.SmoothFamilyTorusInverse.JetBound f S n C = ∀ j ≤ n, ∀ p ∈ S, ∀ (Y : NavierStokes.TorusInverse.Plane), ‖iteratedFDeriv ℝ j f (p, Y)‖ ≤ C
Instances For
Multiplier jet constant used in smooth family torus inverse.
Equations
- One or more equations did not get rendered due to their size.
- NavierStokes.SmoothFamilyTorusInverse.multiplierJetConstant 0 x✝ = 3 ^ (x✝ + 4) * ∑' (k : NavierStokes.TorusInverse.Frequency), (NavierStokes.TorusInverse.weight k ^ 4)⁻¹
Instances For
The loss l+4 comes only from the order-l multiplier and the summable two-dimensional lattice majorant; it does not grow with jet order.
Uniform full-tensor bound with five torus derivatives lost. The constant is chosen before the source, parameter set, or input bound.
Parameter jet, given by (iteratedFDeriv ℝ q f z).compContinuousLinearMap (fun _ => ContinuousLinearMap.inl ℝ P Plane).
Equations
- NavierStokes.SmoothFamilyTorusInverse.parameterJet q f z = (iteratedFDeriv ℝ q f z).compContinuousLinearMap fun (x : Fin q) => ContinuousLinearMap.inl ℝ P NavierStokes.TorusInverse.Plane
Instances For
Parameter jet apply, defined pointwise by parameterJet q f z v.
Equations
Instances For
Nonzero multiplier, with branches according to k = 0.
Instances For
Mixed jet, given by tensorTorusWord w (fun Y => parameterJet q f (z.1, Y)) z.2.
Equations
- One or more equations did not get rendered due to their size.
Instances For
A finite-loss estimate for an actual operator propagates through the successive slow derivatives used in radial integration by parts. This lemma is instantiated below with the constructed torus inverse.
The exact defect in the compact transport primitive, with all auxiliary
slow variables retained in E.
Equations
- NavierStokes.UniformFourierAlias.exactAlias χ M v f z = deriv χ z.1 • NavierStokes.TransportPrimitive.totalIntegral M v f z
Instances For
Repeated integration by parts for full derivative tensors. The last premise bounds genuine source jets, not the alias.
A source-independent constant for the compactification defect, retaining the entire cutoff Leibniz expansion.
Complexify, given by f z.
Equations
- NavierStokes.UniformFourierAlias.complexify f z = ↑(f z)
Instances For
Reassociate radial, slow, and torus variables without changing the norm.
Instances For
From product, given by f ((z.1, z.2.1), z.2.2).
Instances For
Source mean, given by FourierAlias.torusMean (fun Y => f (p.1, (p.2, Y))).
Equations
- NavierStokes.UniformFourierAlias.sourceMean f p = NavierStokes.FourierAlias.torusMean fun (Y : NavierStokes.TorusInverse.Plane) => f (p.1, p.2, Y)
Instances For
Source periodic, given by ∀ U s, FourierAlias.TorusPeriodic (fun Y => f (U, (s, Y))).
Equations
- NavierStokes.UniformFourierAlias.SourcePeriodic f = ∀ (U : ℝ) (s : S), NavierStokes.FourierAlias.TorusPeriodic fun (Y : NavierStokes.TorusInverse.Plane) => f (U, s, Y)
Instances For
Radial slice, given by f (z.1, (s, z.2)).
Instances For
Parameter periodic, given by ∀ p, FourierAlias.TorusPeriodic (fun Y => f (p, Y)).
Equations
- NavierStokes.UniformFourierAlias.ParameterPeriodic f = ∀ (p : P), NavierStokes.FourierAlias.TorusPeriodic fun (Y : NavierStokes.TorusInverse.Plane) => f (p, Y)
Instances For
Parameter mean, given by FourierAlias.torusMean (fun Y => f (p, Y)).
Equations
Instances For
A genuine real directional inverse, obtained from the actual complex Fourier inverse by real part.
Equations
Instances For
Real centered, given by f z - parameterMean f z.1.
Equations
Instances For
Family inverse, given by fromProduct (SmoothFamilyTorusInverse.inverse d (toProduct f)).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Real center source, given by fromProduct (realCentered (toProduct f)).
Equations
Instances For
Admissible, constructed using ContDiff.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Actual, source-uniform finite-seminorm estimate for the translated total integral. Every integration by parts consumes one radial derivative and five orders for the genuine torus inverse.
The retained alias has the same arbitrary inverse-frequency gain, with all source dependence confined to a finite actual derivative bound.
The flat radial weight absorbs every fixed inverse-edge power, giving a global finite seminorm bound for a genuine mean-class family.
The actual power-chart normalization transports the source family class; all constants are uniform in the band and every auxiliary variable.
A proved finite loss on each parameter fiber preserves the actual all-order mean class, including its radial weight.
The actual slow scale, clipped only at the finitely many initial bands.
Equations
Instances For
Choose the number of integrations by parts before learning the finite polynomial slow-scale loss of the required source seminorm.
The concrete radial strip with the manuscript's actual band scales.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Band-dependent sources in the actual all-order mean class have uniformly superflat aliases. The proof chooses one finite IBP order before extracting the required source seminorm and its polynomial growth degree.
Only the integrated torus mean must vanish. Mean subtraction costs four extra finite derivative orders and changes the exact alias by zero.