Branch-and-Cut Solvers for Mixed-Integer Bilevel Linear Programs - Part 1/2

[@glongoria8004]

Great presentation. I'm a bit confused by the bilevel-free set S here 1:07:00, the theorem box, at the bottom, says "does not contain any bilevel-feasible point (not even on its frontier)" I'm wondering if there is a typo and should say INfeasible instead of feasible, like so, "... any bilevel-infeasible point..." ?

[@matheusota5138]

I think there is no typo there. In order for the intersection cut to be valid, the set should not contain bilevel-feasible points (indeed the constraint d^T y > d^T \hat{y} enforce that, because then S(\hat{y}) has no optimum solutions for the follower subproblem).

[@kanezhang4599]

Thanks for the valuable information.