An Approach to Lagrangian Duality in Vector Optimization

Abstract

For a vector extremum problem, optimality conditions, regularity and duality can be studied by means of vector-valued weak separation functions; in this paper duality theorems are established for a wide class of multiobjective problems, by choosing a suitable class of vector-valued weak separation functions which is strictly related to a generalized vector-valued Lagrangian function. The suggested approach allows us to obtain some regulatiry conditions.