Originally posted by: dbaena

Hi,

 

I have a doubt:

 

I have a primal formulation, I solve it with the cplex.setParam(IloCplex::RootAlg,IloCplex::Dual).  Sometime, I hope to obtain an unbounded dual (infeasible prima) because I use dual algorithm, however I obtain a infeasible solution. So, I use dual algoritm but I obtain the solution for primal problem!!

 

Can I use Dual (algorithm and solution) using cplex with primal formulation?

 

For example, I have:

min 0

s.to x>=12

        x <= 10

 

I obtain Infeasible problem, however I want to obtain UNBOUNDED solution because I am using dualSimplex..

 

It's possible?

 

Thanks a lot!!!

 

 

 

 

 

#CPLEXOptimizers
#DecisionOptimization

The RootAlgorithm parameter selects the algorithm for solving the root node (primal) LP. Using the dual simplex algorithm to solve the LP does not change the fact that you are solving the primal LP. In other words, the dual simplex method operates on the same LP that primal simplex does.

In almost all cases, an infeasible primal LP implies an unbounded dual LP. (They can both be infeasible, but that is rare.) Assuming that you are using one of the programming APIs, you can use the IloCplex.isDualFeasible() method (Java name) to confirm that your infeasible primal has an unbounded dual. If you need a dual ray, the IloCplex.dualFarkas() method will provide one.

#CPLEXOptimizers
#DecisionOptimization