Tangent cone #
In this file, we define two predicates UniqueDiffWithinAt 𝕜 s x and UniqueDiffOn 𝕜 s
ensuring that, if a function has two derivatives, then they have to coincide. As a direct
definition of this fact (quantifying on all target types and all functions) would depend on
universes, we use a more intrinsic definition: if all the possible tangent directions to the set
s at the point x span a dense subset of the whole subset, it is easy to check that the
derivative has to be unique.
Therefore, we introduce the set of all tangent directions, named tangentConeAt,
and express UniqueDiffWithinAt and UniqueDiffOn in terms of it.
One should however think of this definition as an implementation detail: the only reason to
introduce the predicates UniqueDiffWithinAt and UniqueDiffOn is to ensure the uniqueness
of the derivative. This is why their names reflect their uses, and not how they are defined.
Implementation details #
Note that this file is imported by Mathlib/Analysis/Calculus/FDeriv/Basic.lean. Hence, derivatives
are not defined yet. The property of uniqueness of the derivative is therefore proved in
Mathlib/Analysis/Calculus/FDeriv/Basic.lean, but based on the properties of the tangent cone we
prove here.
The set of all tangent directions to the set s at the point x.
A point y belongs to the tangent cone of s at x iff
there exist a family of scalars c n, a family of vectors d n,
and a nontrivial filter in the index type such that
d n → 0along the filter;x + d n ∈ seventually along the filter;c n • d n → yalong the filter,
The actual definition is given in terms of cluster points of a filter,
see mem_tangentConeAt_of_seq and exists_fun_of_mem_tangentConeAt
for the two implications unfolding this definition in more convenient way.
In a space with first countable topology,
one can assume that the index type is ℕ and the filter is atTop,
but the definition we use is more useful without that assumption.
Equations
Instances For
Let c n be a family of scalars, d n be a family of vectors, and l be a filter such that
d n → 0alongl;x + d n ∈ sfrequently alongl;c n • d n → yalongl.
Then y belongs to the tangent cone of s at x.
See also
mem_tangentConeAt_of_seqfor a version assuming thatx + d n ∈ seventually alongl.exists_fun_of_mem_tangentConeAtfor the other implication.
A special case of mem_tangentConeAt_of_frequently, which avoids Filter.Frequently.
If y belongs to the tangent cone of s at x, then there exist
- an index type
αand a nontrivial filterlonα; - a family of scalars
c n,n : α, and a family of vectorsd n,n : αsuch that d n → 0alongl;x + d n ∈ seventually alongl;c n • d n → yalongl.
In fact, one can take α = R × E, c = Prod.fst, and d = Prod.snd, but this is not important,
so the lemma statement hides these details.
This lemma provides a convenient way to unfold the definition of tangentConeAt.
"Positive" tangent cone to s at x.
Equations
- posTangentConeAt s x = tangentConeAt NNReal s x
Instances For
A property ensuring that the tangent cone to s at x spans a dense subset of the whole space.
The main role of this property is to ensure that the differential within s at x is unique,
hence this name. The uniqueness it asserts is proved in UniqueDiffWithinAt.eq in
Mathlib/Analysis/Calculus/FDeriv/Basic.lean.
To avoid pathologies in dimension 0, we also require that x belongs to the closure of s (which
is automatic when E is not 0-dimensional).
- dense_tangentConeAt : Dense ↑(Submodule.span R (tangentConeAt R s x))
Instances For
A property ensuring that the tangent cone to s at any of its points spans a dense subset of
the whole space. The main role of this property is to ensure that the differential along s is
unique, hence this name. The uniqueness it asserts is proved in UniqueDiffOn.eq in
Mathlib/Analysis/Calculus/FDeriv/Basic.lean.
Equations
- UniqueDiffOn R s = ∀ x ∈ s, UniqueDiffWithinAt R s x