Please, I have questions about how Gurobi computes the IIS for MIP models.
As mentioned in this article: How does Gurobi compute the IIS for infeasible models? – Gurobi Support Portal. The approach for computing IIS is based on this paper:
Gleeson and Ryan use a variant of Farkas’Lemma to obtain a polyhedron in which each vertex corresponds to an IIS. From my understanding, this method works well for LP problems, and I think it is actually possible to find all IISes of infeasible LP problems if we find all the vertices of the alternative polyhedron.
-Why Gurobi doesn't return all IISes for LP problem?
-How does Gurobi use Gleeson and Ryan's approach, which is for LP, to compute the IIS for MIP (infeasible models with integer variables)?
Thank you very much for your time and consideration
Please sign in to leave a comment.