Hello, I just tried the same problem with the same command line arguments with version 4.53 and I'm getting the same problem where the solution file contains the line:
Status: INTEGER UNDEFINED but that status is wrong. It should be something like NO FEASIBLE SOLUTION or NO INTEGER FEASIBLE SOLUTION so that my third party library, pulp in this case, can return the proper status to my application. Was this bug supposed to be solved in 4.53 or should we wait for an other release for this? Regards, 2013-09-11 14:53 GMT-04:00 Andrew Makhorin <[email protected]>: > > > This may happen when infeasibility is detected by the mip preprocessor > > > (not by the mip solver), which _erroneously_ does not change the mip > > > solution status, so it remains undefined. In any case, if you see on > the > > > terminal something like follows: > > > > > > Preprocessing... > > > PROBLEM HAS NO PRIMAL FEASIBLE SOLUTION > > > > > > it means that no integer feasible solution exists. > > > > > I'm using GLPK with the third party library pulp. Does it means that my > > third party library should scan the stdout in addition to the solution > > file in order to determine the true status of the solution? > > > > No way. This is a glpk bug; it will be fixed in a next release. > > >
_______________________________________________ Help-glpk mailing list [email protected] https://lists.gnu.org/mailman/listinfo/help-glpk
