Thomas Opfer

Let's w.l.o.g. rewrite your problem to the following standard form: min 100 x + ys.t.x + y >= 1- x >= - 1- y >= - 1x, y >= 0 Then the dual problem is: max d1 - d2 - d3s.t.d1 - d2 <= 100d1 - d3 <= 1...

Hi Andy, an equality constraint has kind of a slack variable with an upper and lower bound of 0. When the equality constraint is basic, then this variable is part of the basis. Best regards,Thomas

Without having verified this, I guess you are using the default lower bound of 0. You should add the variables with lower bound -INF. Best regards,Thomas

As I said, please put the libraries at the end of the command. Please try g++ -I/media/34GB/Arquivos-de-Programas-Linux/gurobi801/linux64/include/ -L/media/34GB/Arquivos-de-Programas-Linux/gurobi8...

If you know bounds for every variable that cannot be exceeded, you could add them to model to resolve the ambiguity between infeasibility and unboundedness. If every variable is bounded, then the m...

I guess, I would prefer the 8 cores version then, but for each individual MILP model, it is hard to predict what makes more sense. Maybe, the Gurobi people can provide some benchmarks.

Maybe, there are better ways to achieve your goal, but couldn't you call terminate(), then remove the objective function and then restart primal simplex. This should terminate in 0 (1?) steps as th...

Dear Robert, both 9.70E-12 and 1.70E-10 are to be considered integer (zero)! Please accept this or re-read the article Greg mentioned. Besides of this, you should not use huge numbers like 99999999...

Whether your approach makes sense or not depends on the LP algorithm. Please keep in mind that in dual simplex, the first feasible solution that is found typically is already optimal and early term...

I think NumNZs is what you are looking for, see here: https://www.gurobi.com/documentation/8.1/refman/numnzs.html#attr:NumNZs