劉正傳J. C. Liu2016-04-282016-04-281993-07-??http://rportal.lib.ntnu.edu.tw/handle/20.500.12235/77689Necessary and sufficient conditions without constraint qualification for pareto optimality of multiobjective programming are derived. This article suggests the establishment of a Wolfe-type duality theorem for nondifferentiable, convex, multiobjective minimization problems. The vector Lagrangian and the generalized saddle point for pareto optimality are studied. Key works. Multiobjective programming, pareto optimality, cone of directions, vector Lagrangian, Pareto saddle point.