2024-11-25

874: For C Vectors Bundle and Section from Subset of Base Space Ck at Point Where 0<k, There Is Ck Extension on Open-Neighborhood-of-Point Domain

<The previous article in this series | The table of contents of this series | The next article in this series>

description/proof of that for C vectors bundle and section from subset of base space Ck at point where 0<k, there is Ck extension on open-neighborhood-of-point domain

Topics


About: C manifold

The table of contents of this article


Starting Context



Target Context


  • The reader will have a description and a proof of the proposition that for any C vectors bundle and any section from any subset of the base space Ck at any point where 0<k, there is a Ck extension on an open-neighborhood-of-point domain.

Orientation


There is a list of definitions discussed so far in this site.

There is a list of propositions discussed so far in this site.


Main Body


1: Structured Description


Here is the rules of Structured Description.

Entities:
(E,M,π): { the C vectors bundles of rank k}
S: M
p: S
s: :Sπ1(S)E, { the sections of π|π1(S)}, Cl at p, where 0<l
//

Statements:
VpM{ the open neighborhoods of p},s:Vpπ1(Vp){ the sections of π|π1(Vp)}(s|SVp=s|SVps{ the Cl maps })
//


2: Note


Compare with the proposition that for any map from any subset of any C manifold with boundary into any subset of any C manifold Ck at any point, there is a Ck extension on an open-neighborhood-of-the-point domain, which requires the codomain to be a subset of a C manifold without boundary. An issue because of which that proposition cannot be directly applied is that E may be with a nonempty boundary (when M is with a nonempty boundary), and another issue is that the extension is required to be a section, which that proposition does not guarantee. But the basic idea of this proposition is the same with that of that proposition: the rough reason for this proposition is that while E is locally Up×Rk, the boundary can exist only in the Up part, but the extension needs to be the identity map with respect to the Up part (because it is a section), and the concern is really about only the Rk part, which has no boundary.


3: Proof


Whole Strategy: Step 1: take a chart trivializing subset of M around p, with the chart, (UpM,ϕp), and the induced chart, (π1(Up)E,ϕp~); Step 2: for the components function, f:=ϕp~sϕp1|ϕp(UpS):ϕp(UpS)ϕp~(π1(Up)), take an open neighborhood of ϕp(p), Uϕp(p)Rd, and an extension of f, f:Uϕp(p)Rd+k; Step 3: tweak f to have f:Uϕp(p)Rk×ϕp(Up); Step 4: take Vp:=ϕp1(ϕp(Up)Uϕp(p)) and s:=ϕp~1fϕp|Vpπ1(Vp); Step 5: see that s satisfies the requirements.

Step 1:

Let us take a chart trivializing subset of M around p, with the chart, (UpM,ϕp), which is possible by the proposition that for any C vectors bundle, there is a chart trivializing open cover.

Let us take the induced chart, (π1(Up)E,ϕp~), which is possible by the proposition that for any C vectors bundle, the trivialization of any chart trivializing open subset induces the canonical chart map.

s(Up)π1(Up), because s is a section.

Step 2:

Let the components function be f:=ϕp~sϕp1|ϕp(UpS):ϕp(UpS)ϕp~(π1(Up)).

There is an open neighborhood of ϕp(p), Uϕp(p)Rd, and a Cl extension of f, f:Uϕp(p)Rd+k, because that is what the definition of map between arbitrary subsets of C manifolds with boundary Ck at point, where k excludes 0 and includes requires.

Step 3:

Denote f:(x1,...,xd)(f1(x1,...,xd),...,fk(x1,...,xd),fk+1(x1,...,xd),...,,fd+k(x1,...,xd)). Each fj(x1,...,xd) is Cl.

Let us define f:Uϕp(p)Rk×Uϕp(p),(x1,...,xd)(f1(x1,...,xd),...,fk(x1,...,xd),x1,...,,xd).

f is obviously Cl.

f|ϕp(UpS)=f|ϕp(UpS)=f|ϕp(UpS), because f|ϕp(UpS)(x1,...,xd)=f|ϕp(UpS)(x1,...,xd)=(f1(x1,...,xd),...,fk(x1,...,xd),fk+1(x1,...,xd),...,,fd+k(x1,...,xd))=(f1(x1,...,xd),...,fk(x1,...,xd),x1,...,,xd)=f|ϕp(UpS)(x1,...,xd), because s is fiber-preserving and ϕp~ is induced from ϕp.

So, f is a Cl extension of f.

Step 4:

ϕp(Up)Uϕp(p)ϕp(Up) is an open neighborhood of ϕp(p) on ϕp(Up).

Let us define Vp:=ϕp1(ϕp(Up)Uϕp(p))M such that VpUp, which is an open neighborhood of p.

Let us define s:=ϕp~1fϕp|Vp:Vpπ1(Vp), which is possible because f|ϕp|Vp(Vp) is into Rk×(Uϕp(p)ϕp(Up))Rk×ϕp(Up)=ϕp~(π1(Up)).

Step 5:

s is indeed a section, obviously.

s is Cl, because while ϕp|Vp is C as :Vpϕp(Vp)Rd or Hd, f is Cl as :Uϕp(p)RdRd+k or Hd+k, and ϕp~1 is C as :Rk×ϕp(Up)Rd+k or Hd+kπ1(Up)E, the only concern for s to be a legitimate chain of Cl maps is that Rd or Hd for the codomain of ϕp|Vp is different from Rd for the domain of f, but ϕp|Vp can be regarded to be ϕp|Vp:Vpϕp(Vp)Rd, which obviously does not change C-ness, so, s is indeed a legitimate chain of Cl maps, and is Cl, by the proposition that for any maps between any arbitrary subsets of any C manifolds with boundary Ck at corresponding points, where k includes , the composition is Ck at the point.

s|VpS=s|VpS, because s|VpS=ϕp~1fϕp|Vp|VpS=ϕp~1fϕp|Vp|VpS=ϕp~1ϕp~sϕp1|ϕp(UpS)ϕp|Vp|VpS=s|Vp|VpS=s|VpS.


References


<The previous article in this series | The table of contents of this series | The next article in this series>