多目标不可微规划的最优性条件

THE OPTIMALITY CONDITIONS IN MULTIOBJECTIVE NONSMOOTH PROGRAMMING

本文利用Clarke广义Jacobi短阵的概念,研究了n维欧氏空间中几类多目标不可微规划,并得到相应的最优性条件和Kuhn-Tucker定理。在研究方法上,将目标函数和约束条件结合起来,作为一个整体统一处理,并采用了经典的罚函数方法。

In this paper, the optimality conditions in multiobjective nonsmooth programming with gen- eral, equality, both equality and inequality constraints are studied, where the objective and con- straint functions are all locally Lipschits functions. The problem is studied in such a way that the objective and constraint functions are combined together and that the classical penalty function method is used.