Re: Can NDSolve find the other solution???

*To*: mathgroup at smc.vnet.net*Subject*: [mg31371] Re: Can NDSolve find the other solution???*From*: "cosmicstring" <cosmicstring at yahoo.com>*Date*: Wed, 31 Oct 2001 03:31:02 -0500 (EST)*References*: <9rdgqi$7q0$1@smc.vnet.net> <9rlsjd$ncg$1@smc.vnet.net>*Sender*: owner-wri-mathgroup at wolfram.com

Thank you... I think my eq.s are stiff so I am trying some other ways with Mathematica (Gear method of Mathematica failed but I have to do some other tests to my program) and meanwhile I will be writing a Fortran code to solve the system. I hope everything will be clearer soon... Thank you again for your interest. "Alois Steindl" <Alois.Steindl+e325 at tuwien.ac.at> wrote in message news:9rlsjd$ncg$1 at smc.vnet.net... > "cosmicstring" <cosmicstring at yahoo.com> writes: > > > When I try NDSolve with those four first order ode's and four initial > > conditions I get only one solution but when I work the system out(without > > transforming to four 1st order eq.s) analytically I get two solutions which > > is trivial because I am solving 2nd order eq.s. > > > Hello, > that's not trivial at all. I would bet that you made some serious > mistake. The first order system and the original system are > equivalent; any solution of one system is also a solution of the other > one. > And the solutions for well-behaved initial value problems are usually > unique, so you get the result you should. > > Mathematica also allows you to use higher order equations, but that will not > help you with this problem. > You will have to re-think your equations and calculations. > > > Now, I would like to know if there is a way to obtain the other solution > > using Mathematica. I need this solution because it is the physical one! > > > The non-physical solution is very likely generated by mistakes in your > setup. > > If you can't find your mistake, you could try to tell us your > equations and the solution you expect. > > Alois > >