非线性规划的拟罚函数—强次可行方向法

Quasi-Penalty Function-Strongly Subfeasible Directions Methods for Nonlinear Programming

本文首先提出非线性规划的拟Ｋｕｈｎ—Ｔｕｃｋｅｒ点和拟罚函数法的概念和思想，然后结合强次可行方向法思想给出问题的两个新型算法，称之为拟罚函数—强次可行方向法．证明了该算法收敛到原问题的拟Ｋｕｈｎ—Ｔｕｃｋｅｒ点．

The new concept and idea of quasi-Kuhn-Tucker point and quasi-Penalty function method are put forward in this paper,then two new algorithms are presented by combining the idea of strongly subfeasible directions method,they are called quasi-Penalty functionstrongly sub feasible directions methods.TIt two algotithm are proved to converge to the primal problem's quasi-kuhn-Tucker point.