Ported for Lean Pool: imports relocated and code adapted to Lean/Mathlib v4.34.0.
Compact selection lemmas for stability arguments (LocallyUniformApproximation) #
Part of the compact-extremum selection tools used in viscosity stability
proofs. Split from Selection.lean; see the umbrella module docstring.
Under locally uniform convergence, selected compact maximum contacts whose points and jets converge give tail-closure membership for the limiting superjet triple.
Under locally uniform convergence, selected compact minimum contacts whose points and jets converge give tail-closure membership for the limiting subjet triple.
Let K ⊆ C. Suppose uᵢ converges locally uniformly to u on K,
u is continuous within K at x, x ∈ K, xᵢ → x, and Jᵢ → J along l.
Assume that the set of indices i for which xᵢ i ∈ K,
IsMaxOn (fun y => uᵢ i y - φ y) K (xᵢ i), and
HasSecondOrderExpansionWithin K φ (xᵢ i) (Jᵢ i) all hold belongs to l.
Assume also that the set of indices i for which
K ∈ nhdsWithin (xᵢ i) C belongs to l. Then ((x, u x), J) belongs to the
tail closure of the superjet graphs on C.
Let K ⊆ C. Suppose uᵢ converges locally uniformly to u on K,
u is continuous within K at x, x ∈ K, xᵢ → x, and Jᵢ → J along l.
Assume that the set of indices i for which xᵢ i ∈ K,
IsMinOn (fun y => uᵢ i y - φ y) K (xᵢ i), and
HasSecondOrderExpansionWithin K φ (xᵢ i) (Jᵢ i) all hold belongs to l.
Assume also that the set of indices i for which
K ∈ nhdsWithin (xᵢ i) C belongs to l. Then ((x, u x), J) belongs to the
tail closure of the subjet graphs on C.
Selected compact maximum contacts on a compact relative neighborhood give tail-closure membership over the ambient domain.
Selected compact minimum contacts on a compact relative neighborhood give tail-closure membership over the ambient domain.
Assume K is nonempty and compact, and φ is continuous on K. If the set of
indices i for which uᵢ i is upper semicontinuous on K belongs to l,
then the set of indices i for which there exists x ∈ K satisfying
IsMaxOn (fun y => uᵢ i y - φ y) K x belongs to l.
A strict compact maximizer of u - φ attracts compact maximizers of
uᵢ - φ under uniform convergence on the compact set.
The strictness hypothesis is deliberately local-neighborhood shaped: outside
each neighborhood of x₀, the value of u - φ on K is separated below its
value at x₀ by a positive gap.
A strict compact minimizer of u - φ attracts compact minimizers of
uᵢ - φ under uniform convergence on the compact set.
The strictness hypothesis is deliberately local-neighborhood shaped: outside
each neighborhood of x₀, the value of u - φ on K is separated above its
value at x₀ by a positive gap.
Compact relative neighborhood version of the strict maximum selection theorem.
The hypothesis hJ says that the jet assigned to the test function varies
continuously at the contact point, and hφexp says this assigned jet is a
second-order expansion of the test function at every selected point.
Compact relative neighborhood version of the strict minimum selection theorem.
The hypothesis hJ says that the jet assigned to the test function varies
continuously at the contact point, and hφexp says this assigned jet is a
second-order expansion of the test function at every selected point.
Locally uniform approximation theorem with a positive identity-Hessian perturbation for superjets.
If J = (p, X) belongs to Superjet C u x, δ > 0, and uᵢ converges
locally uniformly to u on C, then ((x, u x), (p, X + δ I)) belongs to
TailClosureSuperjetGraph C uᵢ l. Equivalently, for every set of indices A
with A ∈ l, this graph point lies in the closure of
⋃ i ∈ A, SuperjetGraph C (uᵢ i).
Locally uniform approximation theorem with a negative identity-Hessian perturbation for subjets.
If J = (p, X) belongs to Subjet C u x, δ > 0, and uᵢ converges locally
uniformly to u on C, then ((x, u x), (p, X - δ I)) belongs to
TailClosureSubjetGraph C uᵢ l. Equivalently, for every set of indices A
with A ∈ l, this graph point lies in the closure of
⋃ i ∈ A, SubjetGraph C (uᵢ i).
Locally uniform approximation theorem for superjets, in tail-closure graph form.
If J ∈ Superjet C u x and uᵢ converges locally uniformly to u on C,
then ((x, u x), J) ∈ TailClosureSuperjetGraph C uᵢ l. Equivalently, for
every set of indices A with A ∈ l, this graph point lies in the closure of
⋃ i ∈ A, SuperjetGraph C (uᵢ i).
Locally uniform approximation theorem for subjets, in tail-closure graph form.
If J ∈ Subjet C u x and uᵢ converges locally uniformly to u on C, then
((x, u x), J) ∈ TailClosureSubjetGraph C uᵢ l. Equivalently, for every set
of indices A with A ∈ l, this graph point lies in the closure of
⋃ i ∈ A, SubjetGraph C (uᵢ i).