A smoothing heuristic for the global optimization of a class of bilinear bilevel programs

We provide a global heuristic for a class of bilevel programs with bilinear objectives and linear contraints. The algorithm relies on an interior point approach that can be seen as a combination of the smoothing and implicit programming techniques wellknown in the MPEC literature. But contrary to what is usually done, here the focus is on global optimality. We tested the algorithm on large-scale instances of a network pricing problem, one of the main applications of this model. The preliminary results show that on hard instances, the heuristic is a good alternative to exact solvers based on mixed 0-1 programming formulations.