serious NDSolve bug?
- To: mathgroup at smc.vnet.net
- Subject: [mg44000] serious NDSolve bug?
- From: Artūras Acus <acus at itpa.lt>
- Date: Fri, 17 Oct 2003 05:14:25 -0400 (EDT)
- Organization: Institute of Theoretical Physics and Astronomy
- Sender: owner-wri-mathgroup at wolfram.com
Hi, All
In version 4.1 the following code works as expected.
\[Xi]min = -5; energija = 2*n + 1 /. n -> 1; \[Xi]max = 5;
teisingasKampas =
FindRoot[
(u[\[Xi]] /. First[sprendinys1 =
NDSolve[{Derivative[2][\[Psi]][\[Xi]] == (\[Xi]^2 - energija)*
\[Psi][\[Xi]], Derivative[1][u][\[Xi]] == \[Psi][\[Xi]]^2,
\[Psi][\[Xi]min] == 0,
Derivative[1][\[Psi]][\[Xi]min] == kampas, u[\[Xi]min] == 0},
{\[Psi][\[Xi]], u[\[Xi]]},
{\[Xi], \[Xi]min, \[Xi]max}]] /. \[Xi] -> \[Xi]max) == 1,
{kampas, -0.04, 0.04}]
In 5.0 I get message, that kampas is not a number. Changing FindRoot[ ]
by RootSearch[] (from Ted Ersek package) did not help, therefore I
suspect that there is serious bug in NDSolve argument evaluation
procedure. How I can make it to work ?
Sincerely, Arturas Acus