在黎曼流形上给出了Lipschitz函数的广义方向导数和广义梯度的概念,利用黎曼流形局部上与欧氏空间开集微分同胚的性质以及切映射和余切映射导出了广义梯度的性质和运算法则,证明了定义在黎曼流形上的函数取得极小值的必要条件是广义梯度包含零元素,并利用这些性质给出了黎曼流形上数学规划问题的Fritz John型最优性条件.
The definitions of generalized directional derivative and generalized gradient of Lipschitz functions defined on Riemannian manifold are presented. Some properties of the directional derivative and gradient are proved by using tangent and cotangent mapping. The minimization necessary condition of nonsmooth Lipschitz functions is given. Moreover, Fritz John necessary optimality condition in mathematical programming is provided on Riemannian manifold.