Re: Flow of a differential equation
- From: "W. Dale Hall" <wd_hall@xxxxxxxxxxx>
- Date: Tue, 16 Oct 2007 01:41:15 GMT
Victor B. wrote:
Hello,I'm unsure of what you're assuming and what you wish to prove.
here is a proposition I'm not sure of.
"Let f be a diffeomorphism of R^n. Prove that, given a _continuous_
family of diffeomorphisms (f_t)_t from R^n, with f_0 = Id and f_1=f,
f_t is the flow of a certain differential system".
Is this true ? How can I prove this ?
If you are saying
"give me f : R^n --> R^n a diffeomorphism
and I'll conclude that there is a family
of diffeomorphisms
f_t : R^n --> R^n
(so that the rest follows...)"
then I'd say you're wrong (see explanation
below my .sig).
However, if you're assuming that you are given
a parametrized family {f_t | 0 <= t <= 1} of
self-diffeomorphisms of R^n, for which
f_0 = id_R^n, and f_1 = f,
then here's how to construct a vector field
V on R^n x [0,1], where coordinates of R^n x I
are given as (x,t):
V(x,t) = d/dt (f_t(x))
that is, V(x,t) is the velocity vector of the path
followed by x under the isotopy {f_t | t in [0,1]}.
Note that f_t(x) is, by construction, the solution
of the (time-dependent) DE
X' = V(X,t)
under the initial conditions X(0) = f_0(x).
Why? Because *that's what it means* to have
a solution of the time-dependent DE: a curve
for which the velocity vector at each point
is given by the vector field defining the DE.
That's all I know for now.
Dale
Here's why the first interpretation won't do:
It won't be true if f doesn't preserve orientation.
The diffeomorphism
f: (x1, x2, ..., xn) |--> (-x1, x2, x3, ..., xn)
is not isotopic (i.e. homotopic through homeomorphisms) to
the identity map. To prove it, note that one could take any
isotopy, extend in the obvious way to an isotopy of self-maps
of S^n [and if it ain't obvious: map the point at oo to itself
throughout]. The isotopy in S^n begins with a degree 1 self-
map (namely, the identity) and ends with a degree -1 self-
map. Since degree is a homotopy invariant (it's already a
homology invariant), this can't happen.
.
Thank you very much
Victor
- Follow-Ups:
- Re: Flow of a differential equation
- From: Victor B.
- Re: Flow of a differential equation
- References:
- Flow of a differential equation
- From: Victor B.
- Flow of a differential equation
- Prev by Date: Re: Putnam problem
- Next by Date: Newtons Method---Iteration
- Previous by thread: Flow of a differential equation
- Next by thread: Re: Flow of a differential equation
- Index(es):
Relevant Pages
|