Definición de diferencial
editar
Sean
M
,
N
{\displaystyle M_{}^{},N}
variedades diferenciables ,
F
:
M
⟶
N
{\displaystyle F:M\longrightarrow {}N}
una aplicación diferenciable y
p
∈
M
{\displaystyle p\in M}
, llamaremos diferencial de
F
{\displaystyle F_{}^{}}
a
∀
d
p
F
:
T
p
M
⟶
T
F
(
p
)
N
δ
↦
d
p
F
(
δ
)
:
F
(
M
)
⟶
R
g
↦
d
p
F
(
δ
)
(
g
)
:=
δ
(
g
∘
F
)
.
{\displaystyle {\begin{matrix}{\forall d_{p}F:}&{{\mathcal {T}}_{p}M}&\longrightarrow {}&{{\mathcal {T}}_{F(p)}N}\\&{\delta }&\mapsto &{\begin{matrix}d_{p}F(\delta ):&{\mathcal {F}}(M)&\longrightarrow {}&\mathbb {R} \\&g&\mapsto &d_{p}F(\delta )(g):=\delta (g\circ F).\end{matrix}}\end{matrix}}}
.
Observaciones
Queda claro que
δ
{\displaystyle \delta _{}^{}}
es
δ
p
{\displaystyle \delta _{p}^{}}
, ya que
p
{\displaystyle p_{}^{}}
es redundante pues hablamos de elementos de
T
p
M
{\displaystyle {\mathcal {T}}_{p}M}
y, es decir, derivaciones a
M
{\displaystyle M_{}^{}}
precisamente en
p
{\displaystyle p_{}^{}}
.
Veamos que está bien definida, es decir, que
d
p
F
(
δ
)
∈
T
F
(
p
)
N
{\displaystyle d_{p}F(\delta )\in {\mathcal {T}}_{F(p)}N}
como se ha requerido:
∀
f
,
g
∈
F
(
N
)
,
∀
λ
∈
R
{\displaystyle \forall f,g\in {\mathcal {F}}(N),\forall \lambda \in \mathbb {R} }
,
d
p
F
(
δ
)
(
f
+
g
)
=
δ
(
(
f
+
g
)
∘
F
)
=
δ
(
f
∘
F
+
g
∘
F
)
=
δ
(
f
∘
F
)
+
δ
(
g
∘
F
)
=
d
p
F
(
δ
)
(
f
)
+
d
p
F
(
δ
)
(
g
)
,
{\displaystyle d_{p}^{}F(\delta )(f+g)=\delta ((f+g)\circ F)=\delta (f\circ F+g\circ F)=\delta (f\circ F)+\delta (g\circ F)=d_{p}F(\delta )(f)+d_{p}F(\delta )(g),}
d
p
F
(
δ
)
(
λ
f
)
=
δ
(
(
λ
f
)
∘
F
)
=
δ
(
λ
(
f
∘
F
)
)
=
λ
δ
(
f
∘
F
)
=
λ
d
p
F
(
δ
)
(
f
)
,
{\displaystyle d_{p}^{}F(\delta )(\lambda f)=\delta ((\lambda f)\circ F)=\delta (\lambda (f\circ F))=\lambda \delta (f\circ F)=\lambda d_{p}F(\delta )(f),}
d
p
F
(
δ
)
(
f
⋅
g
)
=
δ
(
(
f
⋅
g
)
∘
F
)
=
δ
(
(
f
∘
F
)
⋅
(
g
∘
F
)
)
=
δ
(
f
∘
F
)
(
g
∘
F
)
(
p
)
+
(
f
∘
F
)
(
p
)
δ
(
g
∘
F
)
=
d
p
F
(
δ
)
(
f
)
g
|
F
(
p
)
+
f
|
F
(
p
)
d
p
F
(
δ
)
(
g
)
{\displaystyle d_{p}^{}F(\delta )(f\cdot g)=\delta ((f\cdot g)\circ F)=\delta ((f\circ F)\cdot (g\circ F))=\delta (f\circ F)(g\circ F)(p)+(f\circ F)(p)\delta (g\circ F)=d_{p}F(\delta )(f)g_{|F(p)}+f_{|F(p)}d_{p}F(\delta )(g)}
,
y, por tanto, es una derivación; en resumen, el diferencial de una derivación es una derivación.
Veamos finalmente que
d
p
F
{\displaystyle d_{p}^{}F}
es
R
{\displaystyle \mathbb {R} }
-lineal:
∀
δ
1
,
δ
2
∈
T
p
M
,
∀
f
∈
F
(
N
)
y
∀
λ
∈
R
{\displaystyle \forall \delta _{1},\delta _{2}\in {\mathcal {T}}_{p}M,\forall f\in {\mathcal {F}}(N)\;y\;\forall \lambda \in \mathbb {R} }
, tenemos
d
p
F
(
δ
1
+
δ
2
)
(
f
)
=
(
δ
1
+
δ
2
)
(
f
∘
F
)
=
δ
1
(
f
∘
F
)
+
δ
2
(
f
∘
F
)
=
d
p
F
(
δ
1
)
(
f
)
+
d
p
F
(
δ
2
)
(
f
)
{\displaystyle d_{p}^{}F(\delta _{1}+\delta _{2})(f)=(\delta _{1}+\delta _{2})(f\circ F)=\delta _{1}(f\circ F)+\delta _{2}(f\circ F)=d_{p}F(\delta _{1})(f)+d_{p}F(\delta _{2})(f)}
,
d
p
F
(
λ
δ
1
)
(
f
)
=
(
λ
δ
1
)
(
f
∘
F
)
=
λ
δ
1
(
f
∘
F
)
=
λ
d
p
F
(
δ
1
)
(
f
)
{\displaystyle d_{p}^{}F(\lambda \delta _{1})(f)=(\lambda \delta _{1})(f\circ F)=\lambda \delta _{1}(f\circ F)=\lambda d_{p}F(\delta _{1})(f)}
,
y por tanto, al ser lineal y bien definida, hereda correctamente las propiedades de suma vectorial y producto por escalar para que los elemento obtenidos en
T
F
(
p
)
N
{\displaystyle {\mathcal {T}}_{F(p)}N}
, a partir de los elementos de
T
p
M
{\displaystyle {\mathcal {T}}_{p}M}
, puedan formar un subespacio vectorial, sería deseable conseguir una base para generar totalmente
T
F
(
p
)
N
{\displaystyle {\mathcal {T}}_{F(p)}N}
.
Así pues, tenemos que
d
p
F
{\displaystyle d_{p}^{}F}
, como aplicación lineal entre espacios vectoriales, queda totalmente determinada por una matriz.
Sean
M
,
N
,
P
{\displaystyle M,\;N,\;P_{}^{}}
variedades diferenciables,
F
:
M
→
N
{\displaystyle F:M\rightarrow N}
,
G
:
N
→
P
{\displaystyle G:N\rightarrow P}
y
p
∈
M
{\displaystyle p\in M}
, entonces tenemos que:
d
p
(
G
∘
F
)
=
d
F
(
p
)
G
∘
d
p
F
{\displaystyle d_{p}(G\circ F)=d_{F(p)}G\circ d_{p}F}
.
Demostración
∀
h
∈
F
(
P
)
,
{\displaystyle \forall h\in {\mathcal {F}}(P),}
∀
δ
∈
T
p
M
{\displaystyle \forall \delta \in {\mathcal {T}}_{p}M}
, sucesivamente por definición:
d
p
(
G
∘
F
)
(
δ
)
(
h
)
=
δ
(
h
∘
G
∘
F
)
=
{\displaystyle d_{p}(G\circ F)(\delta )(h)=\delta (h\circ G\circ F)=}
d
p
F
(
δ
)
(
h
∘
G
)
=
{\displaystyle d_{p}F(\delta )(h\circ G)=}
d
F
(
p
)
G
(
d
p
F
(
δ
)
)
(
h
)
=
{\displaystyle d_{F(p)}G(d_{p}F(\delta ))(h)=}
d
F
(
p
)
G
∘
d
p
F
(
δ
)
(
h
)
{\displaystyle d_{F(p)}G\circ d_{p}F(\delta )(h)}
.