> I'm a brazilian MsC student (so, sorry for my bad English) from UTFPR > University. > First of all, like to say: GLPK + GMPL rules! I'm learning LP/MILP for > an big research project. After some research and various tests with > commercial and free software solvers, we adopted GLPK to our project.
Thank you for your interest in glpk. > For now we use only glpsol at command line to develop models - but we > are using our custom (incipient for now) build of SciTE to integrate > glpsol in an friendly IDE in the WinNT platform. I want to use (and > disseminate) this IDE to teach at this campus in future. When it reach > a minimum stable status i'll put it on the web (bin and code - with a > pre-build glpsol too, if has no objection). > Ok, future is future. > After this self-presentation... yes, i have questions =) > For now, only the crucial: i need to create custom outputs to some > models, and i'm using printf to do this. Ok, all right. Vars and params > looks fine, but i don't know how to print important info - like the > objective value! I've looking at the mail-list archive and found > herea negative answer to var/constraints attributes. My questions is: > (1) it is true for objective value too? I'm re-processing the objective > function after solve to obtain it... And, (2), the answer in the mail > list still true today? If no, there are plans to implement access to > this attributes in future glpsol versions? Currently there is no way to access the objective function value from mathprog. If you have something like: minimize obj: ...some expression...; you might either use the expression in a printf statement, or, better, introduce a variable taking on the objective value: var foo; s.t. bar: foo = ...some expression...; minimize obj: foo; and then use it in printf. _______________________________________________ Help-glpk mailing list [email protected] http://lists.gnu.org/mailman/listinfo/help-glpk
