Re: NDSolve (loop)
- To: mathgroup at smc.vnet.net
- Subject: [mg116953] Re: NDSolve (loop)
- From: Alois Steindl <Alois.Steindl at tuwien.ac.at>
- Date: Sat, 5 Mar 2011 06:08:06 -0500 (EST)
- References: <ikq9ci$8a4$1@smc.vnet.net>
Hello, greg28 <grega.smrkolj at gmail.com> writes: > Hi, > > I would be grateful if you could help me at the following problem. I'm solving a second-order parabolic partial differention equation with NDSolve. The problem is that I have to obtain a solution for varying values of the parameters. This takes a lot of time as I have to increse the minimum number of grid points (for spatial discretization) rather near 1000 to obtain a smooth solution. I wonder whether there's any option to initiate each subsequent NDSolve with the previous solution so as to speed up the algorithm. For small changes in the parameters the solutions are expected not to differ "too much" from each other, so I guess Mathematica could find each subsequent solution easier if it somehow knew the previous solution... Thanks! > I am wondering what you really would like to achieve: You are simulating a family of initial value problems. What is the "message" of the individual solutions? I guess you are looking for some limit, as time t becomes large. In that case I would suggest to directly search for the limiting solution. BTW, by properly formatting your posting (I see a very long line without any breaks) you might increase the number of helpful responses. Alois