Optimality conditions for nonsmooth equilibrium problems via Hadamard directional derivative (Q330326): Difference between revisions

The authors introduce new constraint qualifications for mathematical programs with equilibrium constraints (MPEC) formulated as a mathematical program with complementarity constraints. Since there are several reformulations of MPEC, various stationarity concepts arise. Among these stationary notions, an important role is played by M-stationary ((Mordukhovich)-stationarity, see [\textit{B. S. Mordukhovich}, J. Math. Anal. Appl. 183, No. 1, 250--288 (1994; Zbl 0807.49016)]. The author's approach is based on the lower Hadamard directional derivative. They apply this concept for developing new versions of Fritz-John and Karush-Kuhn-Tucker type of M-stationarity for a nonsmooth MPEC. The authors show that these stationary conditions are first-order necessary optimality conditions for MPEC without smoothness assumptions. They use the lower Hadamard derivative to introduce a (nonsmooth) Abadie-type constraint qualification for MPEC and derive M-stationarity as a necessary optimality condition under this new constraint qualification. Namely, they obtain necessary conditions of Fritz-John and Karush-Kuhn-Tucker type for strong isolated and local minima for a nonsmooth MPEC. Some examples are given to illustrate the necessary optimality conditions. The authors also establish sufficient optimality conditions, in which the standard convexity assumptions are replaced by the pseudoconvexity of the sublevel sets. Namely, they show that two different types of M-stationary are sufficient for a strict global minimum where the objective function and constraints have pseudoconvex sublevel sets.