(1) For an orientable manifold
M
M
M with boundary
∂
M
\partial M
∂M, show that:
∙
\bullet
∙ there is a non-vanishing outer normal vector field
N
N
N on
∂
M
\partial M
∂M;
∙
\bullet
∙ given an orientation n-form
Ω
∈
Λ
n
(
M
)
\Omega\in\Lambda^n(M)
Ω∈Λn(M), the form:
Ω
′
=
(
ι
N
Ω
)
∣
∂
M
,
\Omega'=(\iota_N\Omega)|_{\partial M},
Ω′=(ιNΩ)∣∂M,
is a non-vanishing
(
n
−
1
)
(n-1)
(n−1) form on
∂
M
\partial M
∂M, which gives an induced orientation on
∂
M
\partial M
∂M.
Proof:
Firstly, we need to show a proposition:
For
S
S
S is an immersed hypersurface in
M
M
M, and
N
N
N is a vector field along
S
S
S that is nowhere tangent to
S
S
S. Then
S
S
S has a unique orientation. Also, if
ω
\omega
ω is an orientation form for
M
M
M, then
(
ι
N
ω
)
∣
S
(\iota_N\omega)|_S
(ιNω)∣S is an orientation form for
S
S
S.
proof:
Suppose
ω
\omega
ω is an orientation form for
M
M
M, then
σ
=
(
ι
N
ω
)
∣
S
\sigma=(\iota_N\omega)|_S
σ=(ιNω)∣S is an
(
n
−
1
)
(n-1)
(n−1)-form on
S
S
S. It follows that it’s an orientation form for
S
S
S is we can show that it never vanishes. Given any basis
(
E
1
,
…
,
E
n
−
1
)
(E_1,\dots,E_{n-1})
(E1,…,En−1) for
T
p
S
T_pS
TpS, since
N
N
N is nowhere tangent to
S
S
S implies that
(
N
p
,
E
1
,
…
,
E
n
−
1
)
(N_p,E_1,\dots,E_{n-1})
(Np,E1,…,En−1) is a basis for
T
p
M
T_pM
TpM. The fact that
ω
\omega
ω is nonvanishing implies that
σ
p
(
E
1
,
…
,
E
n
−
1
)
=
ω
p
(
N
p
,
E
1
,
…
,
E
n
−
1
)
≠
0.
\sigma_p(E_1,\dots,E_{n-1})=\omega_p(N_p,E_1,\dots,E_{n-1})\neq0.
σp(E1,…,En−1)=ωp(Np,E1,…,En−1)=0.
Thus the orientation determined by
σ
\sigma
σ is the one defined in the statement of the proposition.
Suppose
d
i
m
(
M
)
=
n
dim (M)=n
dim(M)=n, and
ω
\omega
ω is an orientation form for
M
M
M, let
N
N
N be a smooth outward-pointing vector field along
∂
M
\partial M
∂M. The
(
n
−
1
)
(n-1)
(n−1)-form
(
ι
N
ω
)
∣
∂
M
(\iota_N\omega)|_{\partial M}
(ιNω)∣∂M is an orientation form for
∂
M
\partial M
∂M, so
∂
M
\partial M
∂M is orientable.
(2) Show that the Stokes’ theorem is a generalization of the classical Green formula, Gauss formula and Stokes formula in multi-variable calculus.
Proof:
Apply Stokes’ theorem to
1
1
1-form
ω
=
P
d
x
+
Q
d
y
\omega=Pdx+Qdy
ω=Pdx+Qdy, suppose
D
D
D is a compact regular domain in
R
2
\mathbb{R}^2
R2,
P
,
Q
P,Q
P,Q are smooth real-valued functions on
D
D
D, thus we have Green formula.
Similarly, apply it to 2 2 2-form, ω = P d y ∧ d z + Q d z ∧ d x + R d x ∧ d y \omega=Pdy\wedge dz+Qdz\wedge dx+Rdx\wedge dy ω=Pdy∧dz+Qdz∧dx+Rdx∧dy, we have Gauss formula.
And apply it to ω = P d x + Q d y + R d z \omega=Pdx+Qdy+Rdz ω=Pdx+Qdy+Rdz, we have Stokes formula.
(3) Prove that the pull-back of a smooth map
F
:
M
→
N
F:M\rightarrow N
F:M→N defines a linear map:
F
∗
:
H
p
(
N
)
→
H
p
(
M
)
,
F^*:H^p(N)\rightarrow H^p(M),
F∗:Hp(N)→Hp(M),
which commutes with the wedge product.
Proof:
We set
Z
p
(
M
)
=
K
e
r
(
d
:
Ω
p
(
M
)
→
Ω
p
+
1
(
M
)
)
)
=
{
closed p-forms on
M
}
,
\mathcal{Z}^p(M)=Ker(d:\Omega^p(M)\rightarrow \Omega^{p+1}(M)))=\{\text{closed p-forms on }M\},
Zp(M)=Ker(d:Ωp(M)→Ωp+1(M)))={closed p-forms on M},
B
p
(
M
)
=
I
m
(
d
:
Ω
p
−
1
(
M
)
→
Ω
p
(
M
)
)
)
=
{
exact p-forms on
M
}
.
\mathcal{B}^p(M)=Im(d:\Omega^{p-1}(M)\rightarrow \Omega^{p}(M)))=\{\text{exact p-forms on }M\}.
Bp(M)=Im(d:Ωp−1(M)→Ωp(M)))={exact p-forms on M}.
If
ω
\omega
ω is closed, then
d
(
F
∗
ω
)
=
F
∗
(
d
ω
)
=
0
d(F^*\omega)=F^*(d\omega)=0
d(F∗ω)=F∗(dω)=0, so
F
∗
ω
F^*\omega
F∗ω is also closed. If
ω
=
d
η
\omega=d\eta
ω=dη is exact, then
F
∗
ω
=
F
∗
(
d
η
)
=
d
(
F
∗
η
)
F^*\omega=F^*(d\eta)=d(F^*\eta)
F∗ω=F∗(dη)=d(F∗η), which is also exact. Therefore,
F
∗
F^*
F∗ maps
Z
p
(
N
)
\mathcal{Z}^p(N)
Zp(N) into
Z
p
(
M
)
\mathcal{Z}^p(M)
Zp(M) and
B
p
(
N
)
\mathcal{B}^p(N)
Bp(N) into
B
p
(
M
)
\mathcal{B}^p(M)
Bp(M). The induced cohomology map
F
∗
:
H
p
(
N
)
→
H
p
(
M
)
F^*:H^p(N)\rightarrow H^p(M)
F∗:Hp(N)→Hp(M) is defines as follow:
F
∗
[
ω
]
=
[
F
∗
ω
]
.
F^*[\omega]=[F^*\omega].
F∗[ω]=[F∗ω].
If
ω
′
=
ω
+
d
η
\omega'=\omega+d\eta
ω′=ω+dη, then
[
F
∗
ω
′
]
=
[
F
∗
ω
+
d
(
F
∗
η
)
]
=
[
F
∗
ω
]
[F^*\omega']=[F^*\omega+d(F^*\eta)]=[F^*\omega]
[F∗ω′]=[F∗ω+d(F∗η)]=[F∗ω], so this map is well defined.
(4) Show that the 1 1 1-st de Rham cohomology group of S 1 S^1 S1 is H 1 ( S 1 ) = R H^1(S^1)=\mathbb{R} H1(S1)=R.
Proof:
For any orientation form on S 1 S^1 S1 has nonzero integral, thus d i m H 1 ( S 1 ) ≥ 1 dim H^1(S^1)\geq 1 dimH1(S1)≥1. On the other hand, there is an injective linear map from H 1 ( S 1 ) H^1(S^1) H1(S1) into H o m ( π 1 ( S 1 , 1 ) , R ) Hom(\pi_1(S^1,1),\mathbb{R}) Hom(π1(S1,1),R), which is 1 1 1-dimensional, thus H 1 ( S 1 ) H^1(S^1) H1(S1) is dimension 1 1 1, and is spanned by the cohomology class of any orientation form, thus H 1 ( S 1 ) = R H^1(S^1)=\mathbb{R} H1(S1)=R.