In this paper, we partition the index sets of objectives and constraints, and split the original problem into subproblems with a special structure. We obtain the relationships of ε-Pareto point between original problem and subproblems. Moreover, we also discuss the relationships between s -Pareto points of convex multiple objective programming and ε-Pareto saddle points of Lagrange function under the infeasible point partition. Under the given assumptions, we analyze nonconvex programming and derive a #-Pareto saddle point characterization of ε- Pareto points by applying the augmented Lagrange function.
|