home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky sci.math:17460 sci.physics:21816
- Path: sparky!uunet!think.com!news!columbus
- From: columbus@strident.think.com (Michael Weiss)
- Newsgroups: sci.math,sci.physics
- Subject: Re: More on Huygens' principle
- Date: 28 Dec 92 11:45:54
- Organization: Thinking Machines Corporation, Cambridge MA, USA
- Lines: 36
- Message-ID: <COLUMBUS.92Dec28114554@strident.think.com>
- References: <COLUMBUS.92Dec23114933@strident.think.com>
- <kranzer.725258280@adx.adelphi.edu>
- NNTP-Posting-Host: strident.think.com
- In-reply-to: kranzer@adx.adelphi.edu's message of Fri, 25 Dec 1992 04:38:00 GMT
-
- I wrote:
-
- John Baez has posted the outline of two computations that show why Huygens'
- principle holds for the wave equation with an odd number of spatial
- dimensions, except 1, but fails in even dimensions.
-
- I'm a little puzzled about the exception of 1. Isn't u = f(x+t) + g(x-t) a
- general solution to the wave equation in 1 dimension? This appears to
- propagate without leaving "echoes", i.e., if the supports of f and g for
- t=0 are contained in [-a,a], then the support at t>0 is contained in the
- union of [-a-t, a-t] and [-a+t, a+t]. This should work also for
- distributions. What am I missing?
-
- Herbert Kranzer replies:
-
- Close, but no cigar. Imagine two functions f and g with support R such
- that f + g has support in [-a, a]; for example, f(x) = sgn(x+a) and
- g(x) = -sgn(x-a). Then u(x,0) has support in [-a, a], while the support
- of u(x,t) is the entire interval [-a-t, a+t].
-
- Your argument does, in fact, tell half the story. If you look at the
- d'Alembert solution to the initial value problem u(x,0) = v(x),
- u_t(x,0) = w(x) for the one-dimensional wave equation u_tt = u_xx,
- you will find that v(x) propagates as a sharp front as in Huyghens'
- principle, but w(x) leaves echoes.
-
- OK, I think I see. The general solution is
-
- / x+t
- u(x,t) = v(x+t)/2 + v(x-t)/2 + (1/2) | w(s)ds
- / x-t
-
- and v = f+g, w = f'-g'. If f and g *both* have support in [-a,a], then
- v and w cannot be chosen arbitrarily. I suppose a sharp "explosion" at
- t=x=0 is best modelled by taking v identically 0, and w a delta function.
- Thanks.
-