本文利用集值映射的上图导数引进了全局真有效意义下的广义梯度和广义次微分的概念,并且给出了集值映射全局真有效次微分的存在定理,还建立了集值向量优化问题全局真有效解在次微分形式下的最优性条件.
The concepts of the generalized gradients and subdifferential in sense of globally proper efficiency are introduced by epiderivative for a set-valued mapping. The condition of existence for globally proper subdifferential is established. The optimality condition for globally efficient solutions of vector set-valued optimization is obtained hy the subdifferential in terms of globally proper efficiency.