Optimality conditions for approximate solutions in multiobjective optimization problems (Q607984)

This paper presents necessary and sufficient optimality conditions for approximate efficient solutions of multi-objective optimization problems. It uses the tangent cone, the cone of feasible directions and the \(\varepsilon\)-normal cone to derive different characterizations for global approximate solutions in the convex case considered in Section 3. In Section 4, the results are generalized to the non-convex case by using local concepts. First- and second-order conditions for local approximate efficient solutions are derived, which are expressed by means of the tangent cone, the second-order tangent set and the asymptotic second-order set. In addition, some sufficient conditions are presented for local (weakly) approximate solutions by using Hadamard upper (lower) directional derivatives.