|
|
Line 1: |
Line 1: |
− | A mapping of the tangent space of a manifold <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e0369301.png" /> into <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e0369302.png" />. It is defined by a [[Connection|connection]] given on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e0369303.png" /> and is a far-reaching generalization of the ordinary exponential function regarded as a mapping of a straight line into itself.
| + | <!-- |
| + | e0369301.png |
| + | $#A+1 = 82 n = 0 |
| + | $#C+1 = 82 : ~/encyclopedia/old_files/data/E036/E.0306930 Exponential mapping |
| + | Automatically converted into TeX, above some diagnostics. |
| + | Please remove this comment and the {{TEX|auto}} line below, |
| + | if TeX found to be correct. |
| + | --> |
| | | |
− | 1) Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e0369304.png" /> be a <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e0369305.png" />-manifold with an affine connection, let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e0369306.png" /> be a point in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e0369307.png" />, let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e0369308.png" /> be the tangent space to <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e0369309.png" /> at <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693010.png" />, let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693011.png" /> be a non-zero vector in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693012.png" />, and let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693013.png" /> be the geodesic passing through <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693014.png" /> in the direction of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693015.png" />. There is an open neighbourhood <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693016.png" /> of the point <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693017.png" /> in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693018.png" /> and an open neighbourhood <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693019.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693020.png" /> in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693021.png" /> such that the mapping <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693022.png" /> is a diffeomorphism of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693023.png" /> onto <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693024.png" />. This mapping is called the exponential mapping at <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693025.png" /> and is denoted by <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693026.png" />. A neighbourhood <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693027.png" /> is called normal if: 1) the mapping <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693028.png" /> maps <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693029.png" /> onto <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693030.png" /> diffeomorphically; and 2) <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693031.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693032.png" /> imply that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693033.png" />. In this case <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693034.png" /> is said to be a normal neighbourhood of the point <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693035.png" /> in the manifold <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693036.png" />. Every <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693037.png" /> has a convex normal neighbourhood <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693038.png" />: Any two points of such a neighbourhood can be joined by exactly one geodesic segment lying in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693039.png" />. If <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693040.png" /> is a complete Riemannian manifold, then <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693041.png" /> is a surjective mapping of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693042.png" /> onto <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693043.png" />.
| + | {{TEX|auto}} |
| + | {{TEX|done}} |
| | | |
− | 2) Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693044.png" /> be a Lie group with identity <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693045.png" /> and let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693046.png" /> be the corresponding Lie algebra consisting of the tangent vectors to <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693047.png" /> at <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693048.png" />. For every vector <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693049.png" /> there is a unique differentiable homomorphism <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693050.png" /> of the group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693051.png" /> into <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693052.png" /> such that the tangent vector to <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693053.png" /> at <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693054.png" /> coincides with <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693055.png" />. The mapping <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693056.png" /> is called the exponential mapping of the algebra <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693057.png" /> into the group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693058.png" />. There is an open neighbourhood <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693059.png" /> of the point <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693060.png" /> in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693061.png" /> and an open neighbourhood <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693062.png" /> of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693063.png" /> in <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693064.png" /> such that <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693065.png" /> is a diffeomorphism of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693066.png" /> onto <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693067.png" />. Let <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693068.png" /> be some basis for the algebra <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693069.png" />. The mapping <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693070.png" /> is a coordinate system on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693071.png" />; these coordinates are called canonical.
| + | A mapping of the tangent space of a manifold $ M $ |
| + | into $ M $. |
| + | It is defined by a [[Connection|connection]] given on $ M $ |
| + | and is a far-reaching generalization of the ordinary exponential function regarded as a mapping of a straight line into itself. |
| | | |
− | The concept of an exponential mapping of a Lie group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693072.png" /> can also be approached from another point of view. There is a one-to-one correspondence between the set of all affine connections on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693073.png" /> that are invariant relative to the group of left translations and the set of bilinear functions <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693074.png" />. It turns out that the exponential mapping <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693075.png" /> of the algebra <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693076.png" /> into the group <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693077.png" /> coincides with the mapping <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693078.png" /> of the tangent space of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693079.png" /> into the manifold <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693080.png" /> at the point <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693081.png" /> in this manifold with respect to the left-invariant affine connection corresponding to any skew-symmetric bilinear function <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/e/e036/e036930/e03693082.png" />.
| + | 1) Let $ M $ |
| + | be a $ C ^ \infty $- |
| + | manifold with an affine connection, let $ p $ |
| + | be a point in $ M $, |
| + | let $ M _ {p} $ |
| + | be the tangent space to $ M $ |
| + | at $ p $, |
| + | let $ X $ |
| + | be a non-zero vector in $ M _ {p} $, |
| + | and let $ t \rightarrow \gamma _ {X} ( t) $ |
| + | be the geodesic passing through $ p $ |
| + | in the direction of $ X $. |
| + | There is an open neighbourhood $ N _ {0} $ |
| + | of the point $ 0 $ |
| + | in $ M _ {p} $ |
| + | and an open neighbourhood $ N _ {p} $ |
| + | of $ p $ |
| + | in $ M $ |
| + | such that the mapping $ X \rightarrow \gamma _ {X} ( 1) $ |
| + | is a diffeomorphism of $ N _ {0} $ |
| + | onto $ N _ {p} $. |
| + | This mapping is called the exponential mapping at $ p $ |
| + | and is denoted by $ \mathop{\rm exp} $. |
| + | A neighbourhood $ N _ {0} $ |
| + | is called normal if: 1) the mapping $ \mathop{\rm exp} $ |
| + | maps $ N _ {0} $ |
| + | onto $ N _ {p} $ |
| + | diffeomorphically; and 2) $ X \in N _ {0} $ |
| + | and $ 0 \leq t \leq 1 $ |
| + | imply that $ t X \in N _ {0} $. |
| + | In this case $ N _ {p} $ |
| + | is said to be a normal neighbourhood of the point $ p $ |
| + | in the manifold $ M $. |
| + | Every $ p \in M $ |
| + | has a convex normal neighbourhood $ N _ {p} $: |
| + | Any two points of such a neighbourhood can be joined by exactly one geodesic segment lying in $ N _ {p} $. |
| + | If $ M $ |
| + | is a complete Riemannian manifold, then $ \mathop{\rm exp} $ |
| + | is a surjective mapping of $ M _ {p} $ |
| + | onto $ M $. |
| + | |
| + | 2) Let $ G $ |
| + | be a Lie group with identity $ e $ |
| + | and let $ \mathfrak g $ |
| + | be the corresponding Lie algebra consisting of the tangent vectors to $ G $ |
| + | at $ e $. |
| + | For every vector $ X \in \mathfrak g $ |
| + | there is a unique differentiable homomorphism $ \theta $ |
| + | of the group $ \mathbf R $ |
| + | into $ G $ |
| + | such that the tangent vector to $ \theta ( \mathbf R ) $ |
| + | at $ e $ |
| + | coincides with $ X $. |
| + | The mapping $ X \rightarrow \mathop{\rm exp} X = \theta ( 1) $ |
| + | is called the exponential mapping of the algebra $ \mathfrak g $ |
| + | into the group $ G $. |
| + | There is an open neighbourhood $ N _ {0} $ |
| + | of the point $ 0 $ |
| + | in $ \mathfrak g $ |
| + | and an open neighbourhood $ N _ {e} $ |
| + | of $ e $ |
| + | in $ G $ |
| + | such that $ \mathop{\rm exp} $ |
| + | is a diffeomorphism of $ N _ {0} $ |
| + | onto $ N _ {e} $. |
| + | Let $ X _ {1} \dots X _ {n} $ |
| + | be some basis for the algebra $ \mathfrak g $. |
| + | The mapping $ \mathop{\rm exp} ( x _ {1} X _ {1} + {} \dots + x _ {n} X _ {n} ) \rightarrow ( x _ {1} \dots x _ {n} ) $ |
| + | is a coordinate system on $ N _ {e} $; |
| + | these coordinates are called canonical. |
| + | |
| + | The concept of an exponential mapping of a Lie group $ G $ |
| + | can also be approached from another point of view. There is a one-to-one correspondence between the set of all affine connections on $ G $ |
| + | that are invariant relative to the group of left translations and the set of bilinear functions $ \alpha : \mathfrak g \times \mathfrak g \rightarrow \mathfrak g $. |
| + | It turns out that the exponential mapping $ \mathop{\rm exp} $ |
| + | of the algebra $ \mathfrak g $ |
| + | into the group $ G $ |
| + | coincides with the mapping $ \mathop{\rm exp} $ |
| + | of the tangent space of $ \mathfrak g $ |
| + | into the manifold $ G $ |
| + | at the point $ e $ |
| + | in this manifold with respect to the left-invariant affine connection corresponding to any skew-symmetric bilinear function $ \alpha $. |
| | | |
| ====References==== | | ====References==== |
| <table><TR><TD valign="top">[1]</TD> <TD valign="top"> S. Helgason, "Differential geometry, Lie groups, and symmetric spaces" , Acad. Press (1978)</TD></TR></table> | | <table><TR><TD valign="top">[1]</TD> <TD valign="top"> S. Helgason, "Differential geometry, Lie groups, and symmetric spaces" , Acad. Press (1978)</TD></TR></table> |
A mapping of the tangent space of a manifold $ M $
into $ M $.
It is defined by a connection given on $ M $
and is a far-reaching generalization of the ordinary exponential function regarded as a mapping of a straight line into itself.
1) Let $ M $
be a $ C ^ \infty $-
manifold with an affine connection, let $ p $
be a point in $ M $,
let $ M _ {p} $
be the tangent space to $ M $
at $ p $,
let $ X $
be a non-zero vector in $ M _ {p} $,
and let $ t \rightarrow \gamma _ {X} ( t) $
be the geodesic passing through $ p $
in the direction of $ X $.
There is an open neighbourhood $ N _ {0} $
of the point $ 0 $
in $ M _ {p} $
and an open neighbourhood $ N _ {p} $
of $ p $
in $ M $
such that the mapping $ X \rightarrow \gamma _ {X} ( 1) $
is a diffeomorphism of $ N _ {0} $
onto $ N _ {p} $.
This mapping is called the exponential mapping at $ p $
and is denoted by $ \mathop{\rm exp} $.
A neighbourhood $ N _ {0} $
is called normal if: 1) the mapping $ \mathop{\rm exp} $
maps $ N _ {0} $
onto $ N _ {p} $
diffeomorphically; and 2) $ X \in N _ {0} $
and $ 0 \leq t \leq 1 $
imply that $ t X \in N _ {0} $.
In this case $ N _ {p} $
is said to be a normal neighbourhood of the point $ p $
in the manifold $ M $.
Every $ p \in M $
has a convex normal neighbourhood $ N _ {p} $:
Any two points of such a neighbourhood can be joined by exactly one geodesic segment lying in $ N _ {p} $.
If $ M $
is a complete Riemannian manifold, then $ \mathop{\rm exp} $
is a surjective mapping of $ M _ {p} $
onto $ M $.
2) Let $ G $
be a Lie group with identity $ e $
and let $ \mathfrak g $
be the corresponding Lie algebra consisting of the tangent vectors to $ G $
at $ e $.
For every vector $ X \in \mathfrak g $
there is a unique differentiable homomorphism $ \theta $
of the group $ \mathbf R $
into $ G $
such that the tangent vector to $ \theta ( \mathbf R ) $
at $ e $
coincides with $ X $.
The mapping $ X \rightarrow \mathop{\rm exp} X = \theta ( 1) $
is called the exponential mapping of the algebra $ \mathfrak g $
into the group $ G $.
There is an open neighbourhood $ N _ {0} $
of the point $ 0 $
in $ \mathfrak g $
and an open neighbourhood $ N _ {e} $
of $ e $
in $ G $
such that $ \mathop{\rm exp} $
is a diffeomorphism of $ N _ {0} $
onto $ N _ {e} $.
Let $ X _ {1} \dots X _ {n} $
be some basis for the algebra $ \mathfrak g $.
The mapping $ \mathop{\rm exp} ( x _ {1} X _ {1} + {} \dots + x _ {n} X _ {n} ) \rightarrow ( x _ {1} \dots x _ {n} ) $
is a coordinate system on $ N _ {e} $;
these coordinates are called canonical.
The concept of an exponential mapping of a Lie group $ G $
can also be approached from another point of view. There is a one-to-one correspondence between the set of all affine connections on $ G $
that are invariant relative to the group of left translations and the set of bilinear functions $ \alpha : \mathfrak g \times \mathfrak g \rightarrow \mathfrak g $.
It turns out that the exponential mapping $ \mathop{\rm exp} $
of the algebra $ \mathfrak g $
into the group $ G $
coincides with the mapping $ \mathop{\rm exp} $
of the tangent space of $ \mathfrak g $
into the manifold $ G $
at the point $ e $
in this manifold with respect to the left-invariant affine connection corresponding to any skew-symmetric bilinear function $ \alpha $.
References
[1] | S. Helgason, "Differential geometry, Lie groups, and symmetric spaces" , Acad. Press (1978) |