C-efficiency in nondifferentiable vector optimization

We study the minimal solutions in a nondifferentiable multiobjective problem, using a relation induced by a cone C, that is C-efficient and C-weakly efficient solutions. First of all, a new class of nondifferentiable vector functions, named (C1,C2)-pseudoinvex, is introduced pointing out that it differs from the ones already proposed in the literature. Then, it is proved that a critical point is C-efficient or weakly C-efficient if and only if the vector objective function is (C1,C2)-pseudoinvex. The obtained results generalize to the nondifferentiable case some known definitions and characterization theorems which appeared in the recent vector optimization literature.