> I think there is a bug concerning the MILP solver in GLPK version 4.43. > I deactivated the MILP presolver, and explicitly provided a valid basis > by running the simplex method first, followed by glp_intopt: > > retSIMP = glp_simplex(lp, &simplexParams); > retMILP = glp_intopt(lp, &milpParams); > > Occasionally it happens that the MILP solver does not need to branch > (i.e. the MILP callback function is never called with > glp_ios_reason(tree) == GLP_IBRANCH), but an integer solution is found > merely using (Gomory) cutting planes. In these cases, glp_intopt does > not return the integer solution, but the initial, relaxed solution as > found by the simplex method. > I am not sure, if this also happens when the MILP presolver is activated.
Please check that you retrieve mip solution components with glp_mip_row/col_val, not with glp_get_row/col_prim (the latter is used to retrieve components of basic solution). I don't think that this is a bug, because glp_intopt stores to the problem object only integer feasible solutions. Thanks. Andrew Makhorin _______________________________________________ Bug-glpk mailing list [email protected] https://lists.gnu.org/mailman/listinfo/bug-glpk
