引进了集值映射的非导数型超次梯度概念,在某种条件下证明了该梯度的存在性。作为应用,在近似锥次类凸假设下,给出了用该次梯度刻画集值优化问题取得超有效元的充分必要条件。
A type of nonderivative super subgradient was introduced for a set-valued map.Under some condition,its existence theorem was proved.As an application,under the assumption of near-subconvexlikeness,the sufficient and necessary optimality conditions were established for set-valued optimization problem to obtain super efficient elements.