On Fri, 2017-06-16 at 09:07 -0500, Shuo Xu wrote: > well, this is not the case I was trying to talk. > > My problem is that if c and A are highly sparse. > > (I see your source code is using a net-structure to analyze (read > write )the A structure. > > the net you built to analyze the structure of A and c is not a > complete net to re-create the original A and c from LP-file. > > > And, we are trying to use GLPK to write and read LP-files, because we > are trying to make our cases exchangeable between different > researchers.
> And, most solvers are supporting LP file without problem, such as CBC, > LP_solve, Cplex, AMPL, GAMS, and SCIP. > On using other packages you will encounter the same issue, because, as I mentioned earlier, cplex lp format doesn't allow encoding lp/mip in a general form. AFAIK, for example, cplex changes the instance on writing it to a file if it has double-bounded rows. If you need maximum portability, it is better to use the free mps format (it is supported by all packages you named) rather than the cplex lp format. _______________________________________________ Help-glpk mailing list [email protected] https://lists.gnu.org/mailman/listinfo/help-glpk
