摘要
在这篇文章中我们研究了对于不等式约束的非线性规划问题如何根据极小极大问题的鞍点来找精确罚问题的解．对于一个具有不等式约束的非线性规划问题，通过罚函数，我们构造出一个极小极大问题，应用交换“极小”和“极大”次序的策略，证明了罚问题的鞍点定理．研究结果显示极小极大问题的鞍点是精确罚问题的解．
We study in this paper how to find a soution of an exact penalty problem for non- linear programming with inequality constraints by finding the saddle point of a minimax problem. In order to find the saddle points for a minimax problem, we suggest the strat- egy for interchanging the order between min and max. We prove the saddle point theorem for minimax problem. The result shows that any saddle point of minimax problem is the solution of exact penalty problem.
出处
《运筹学学报》
CSCD
北大核心
2001年第1期47-52,共6页
Operations Research Transactions
基金
supported by the Science Foundation of Shanghai Municipal Commission of Education and theNational Natural Science Foundation o
