Off-plane endpoint extensions from actual raw jets #
A compact spacetime localization converts bounds on every actual derivative in a past half-ball into a genuine smooth extension. The physical scale is then localized away from zero, so raw power-logarithmic jet bounds supply the required estimates without a recursive continuation assumption.
A dimension-independent smooth extension of a bounded-jet open strip #
The endpoint values below are derived from bounds on the actual joint
Frechet derivatives, using completeness and the mean-value theorem. The
normal-jet gluing argument is generalized from SpacetimeGluing; no existing
project source is altered and no extension or closed-side regularity is assumed.
Time vector, given by (1, 0).
Instances For
Directional, given by fderivWithin ℝ f s z v.
Equations
- NavierStokes.GenericEndpointExtension.Gluing.directional s f v z = (fderivWithin ℝ f s z) v
Instances For
Normal iter as an element of ℕ → (ℝ × X) → V | 0 => f | n + 1 => directional s (normalIter s f n) timeVector.
Equations
- One or more equations did not get rendered due to their size.
- NavierStokes.GenericEndpointExtension.Gluing.normalIter s f 0 = f
Instances For
Schwarz's theorem commutes two fixed directional derivatives on a regular closed domain; no symmetry of full higher tensors is assumed.
Any fixed directional derivative commutes with every normal iterate.
The joint normal iterates equal the genuine one-dimensional derivatives of the time slice, including at a one-sided boundary.
Matching boundary values gives matching tangential derivatives; together with the first normal derivative this determines the full Frechet derivative.
Matching normal trace functions implies matching normal traces after any directional derivative. The proof derives, rather than assumes, the necessary tangential and mixed derivative equalities.
Actual full Frechet derivatives glue when their boundary values match.
First-order joint gluing needs only value and first normal-derivative matching; spatial derivative matching is a consequence.
Finite-order induction from all matching normal jets. The induction keeps the codomain fixed and differentiates in each spacetime direction.
Joint C∞ gluing, expressed in actual one-sided time-slice jets.
No matching of mixed Frechet tensors is assumed: it is derived from the
normal trace functions and Schwarz's theorem.
The actual normal jet of a closed-past field, viewed as a spatial coefficient for the Taylor--Borel construction.
Equations
Instances For
A constructed global extension: join the closed-past field to the Taylor--Borel realization of its actual normal jets.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Every full mixed jet on the past, including the boundary, is preserved.
Uniform continuity makes the actual values Cauchy at every point of the closure. Completeness supplies a limit; it is not an input assumption.
The genuine joint derivatives are differentiable inside the open set.
A bound for derivative n+1 controls differences of the actual n-th
joint derivative on the convex set.
The continuous completion of an actual joint derivative tensor.
Equations
Instances For
The mean-value theorem identifies the derivatives of the completed tensors on the boundary, including all mixed derivative directions.
Complete the values using the zeroth tensor.
Equations
Instances For
Actual uniform joint derivative bounds give joint smoothness on the closure. In particular no one-sided trace regularity is assumed.
A fiber that vanishes on the open strip also vanishes at its two completed endpoints. This follows from continuity, not a support enlargement.
Every spatial additive period passes to both completed boundary values.
A fixed smooth retraction of the past into the closed strip, equal to
the identity for t ≥ -1/2.
Equations
- NavierStokes.GenericEndpointExtension.lowerClamp t = t * (2 * t + 2).smoothTransition
Instances For
Zero normal coefficients stay zero in the actual Borel series.
Upper closed, given by Gluing.smoothExtension 1 (clamped f) (clamped_contDiffOn hf).
Equations
Instances For
Lower closed, given by reflect (upperClosed (reflect f) (reflect_contDiffOn hf)).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Use the lower continuation for negative parameters and the upper continuation for positive ones. They agree on a whole central strip.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The constructed extension takes only interior smoothness and bounds on the actual joint derivatives as inputs.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Past ball, given by Metric.ball (1, x) r ∩ SpacetimeEndpoint.openPast 1.
Equations
Instances For
Local bump, bundling rIn, rOut, rIn_pos, rIn_lt_rOut.
Equations
Instances For
Localized, given by localBump x hr w • f w.
Equations
- NavierStokes.OffplaneJetExtensions.localized f x hr w = ↑(NavierStokes.OffplaneJetExtensions.localBump x hr) w • f w
Instances For
A local bound for every actual joint derivative supplies an actual smooth extension. Endpoint derivative limits are constructed by the generic extension theorem; they are not additional inputs.
One neighborhood controls the physical scale for all derivative orders. Only the actual past coordinate is used in the conclusion.
Arbitrary real powers and logarithmic losses are uniformly bounded
when the scale lies in a fixed compact subinterval of (0,∞).
A single-field endpoint theorem, including the bounded finite pieces at stage zero. It assumes only their actual interior jet estimates.
Positive raw stages extend directly from RawStageBounds. The stage
index hypothesis is retained exactly: no stage-zero estimate is inferred.
Combine, for example, the existing base-gauge extension with the separately derived extension of the finite initial correction.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The constructed stage need only agree with the bounded model on its actual validity region. The positive scale margin supplies the required past neighborhood automatically.
The finite initialized potential correction has its own derived estimate. This theorem does not apply the positive-stage bound to index zero.