[Date Index] [Thread Index] [Author Index]

Re: NDSolve error message: Can't find starting value ...

• To: mathgroup at smc.vnet.net
• Subject: [mg23786] Re: NDSolve error message: Can't find starting value ...
• From: "Kevin J. McCann" <kevin.mccann at jhuapl.edu>
• Date: Sat, 10 Jun 2000 02:59:31 -0400 (EDT)
• Organization: Johns Hopkins University Applied Physics Lab, Laurel, MD, USA
• References: <8hfggl$i4h@smc.vnet.net>
• Sender: owner-wri-mathgroup at wolfram.com

Axel,

A coulple of things. First off you need to specify R[x] and R'[x] at some
point in the interval to get things moving. Second, I am not sure that a
numerical integration will treat the delta function correctly. You might
consider finding solutions on either side to of x=0 and finding the
constants by integrating over an infinitesimal interval around x=0. In fact,
you are solving for the Green's function of the DE.

Kevin

Axel Kowald <axel at itb.biologie.hu-berlin.de> wrote in message
news:8hfggl$i4h at smc.vnet.net...
> Hello,
>
> I try to solve the following second order ODE with NDSolve
>
> NDSolve[{0 == R''[x] + DiracDelta[x] - R[x],         R''[-3] == 0,
> R''[3] == 0}, R, {x, -3, 3}]
>
> and I get the following error message:
>
> NDSolve::"ndsv": "Cannot find starting value for the variable x."
>
> I couldn't find any description of this message in the Mathematica book,
so I'm
> not sure what to do. I have a second order ODE with two boundary
> conditions, what's missing ?
>
> Many thanks,
>
>                Axel Kowald
>
> P.S. Btw., this is done with Mathematica 4.
>
>
>