Re: a problem about NDSolve and NIntegrate

*To*: mathgroup at smc.vnet.net*Subject*: [mg80200] Re: a problem about NDSolve and NIntegrate*From*: Jean-Marc Gulliet <jeanmarc.gulliet at gmail.com>*Date*: Tue, 14 Aug 2007 07:16:53 -0400 (EDT)*Organization*: The Open University, Milton Keynes, UK*References*: <f9emog$lpq$1@smc.vnet.net> <f9p6g1$s1r$1@smc.vnet.net>

lynn wrote: <snip> > Thanks for your pointing out my faults. > > I miss-typed it when I tried to simplize my equation. > > But my original problem still exist, > > The following is the original equation, > > f[ _?NumberQ] := y[ ] /. NDSolve[{ > 2 y'[ ] y[ ] + > y''[ ] y[ ] - > (y'[ ])^2 == y[ ], y[0.00001] == 1, y'[0.00001] == = > 0}, y, > { , > 0.00001, }, MaxSteps -> 1000][[1]] > > then f[5]=InterpolatingFunction[{{0.00001, 5.}}, <>][ ] > > and I want to integrate it y[ ] with a , > > \!\(g[w_?NumberQ, _?NumberQ] := 2\ > NIntegrate[\ \/y[ ]=D7 \/\@\( \^2 - w\^2\), { , w, }, > MaxRecursion -> \ > 10]\) > > But I can't get correct answer when ask for such as g[1,2]... Your post does not make any sense: do not use any special character set when posting to Usenet. (Or use the Ingolf Dahl's "SetFaceAndFont" palette. See http://library.wolfram.com/infocenter/MathSource/5752/) I believe that f is defined more or less as follows In[1]:= f[c_?NumberQ] := y[x] /. NDSolve[{2 y'[x] y[x] + y''[x] y[x] - (y'[x])^2 == y[x], y[0.00001] == 1, y'[0.00001] == 0}, y, {x , 0.00001, c}, MaxSteps -> 1000][[1]] So you can get an interpolating function with In[2]:= f[5] Out[2]= InterpolatingFunction[][x] and also plot it as in In[3]:= Plot[Evaluate@f[5], {x, 0.00001, 5}] [... graphic deleted ...] As far as I can see in the posted mess, g does not use f, and y[x] is still *NOT* defined outside of f (this have already been pointed out by Jens-Peer Kuska a couple of days ago); therefore g has no knowledge of what could possibly be y[x]. -- Jean-Marc