We address mixed-integer linear bilevel programming. A discussion ofthe relationships between the optimistic and the pessimistic setting ispresented, providing necessary and sufficient conditions for them to beequivalent. A new class of inequalities, the follower optimality cuts, isintroduced and a related single-level non-compact reformulation of theproblem is derived. The same is done for a revision of a family of knowninequalities, the no-good cuts. A polyhedral comparison of the two for-mulations is carried out. Finally, we present a branch-and-cut algorithmbased on the follower optimality cuts and discuss computational results,comparing our algorithm with the best approaches in the literature, toshow its effectiveness.
Citation
working paper 30 dec 2020
Article
View The follower optimality cuts for mixed-integer linear bilevel programming problems