in zeta you will give the locations (relative to y) of your boundary conditions in your case something like [0 0 0 , L L L ] if you have 6 states 3 with initial constraints and 3 with final contraints The constraints should then be defined by the gsub and dgsub functions
Serge Le 12/11/2015 11:09, amonm...@laas.fr a écrit : > Hi everyone, > > I'm trying to understand how bvodeS works and whether I can use it to > solve numerically my problem. > A bit of context: > I have a 1D coupled-mode problem, where I model the propagation along > one direction (let's call it y) of two modes (one in +y-direction and > one in the -y-direction). > Schematically, it looks like this: > > a(0) -->| ODE |-->a(L) > b(0) <--| ODE |<--b(L) > > The ODE is of the first order (only a(y), b(y) and a'(y) and b'(y) in > the ODE, together with non-constant factors). > > The thing is that I have scattering-like boundary conditions. > I know the inputs [a(0)=a0;b(L)=bL] and look for the outputs [a(L)=?; > b(0)=?]. > I am of course also interested in the values of [a(y), b(y)] for all > y in [0,L], but the puzzling issue are my boundary conditions. > I basically half of my conditions on one side, the other half on the > other side. > Can I solve such a system using bvodeS? > If I can, how am I supposed to build "zeta" to reflect that my > conditions only concerns a(0) and b(L)? > Is this doable to get zeta "with holes"? > > Thank you in advance for your help, > > Antoine > _______________________________________________ > users mailing list > users@lists.scilab.org > http://lists.scilab.org/mailman/listinfo/users > _______________________________________________ users mailing list users@lists.scilab.org http://lists.scilab.org/mailman/listinfo/users