摘要
本文利用指数型增广拉格朗日函数将一类广义半无限极大极小问题在一定条件下转化为标准的半无限极大极小问题,使它们具有相同的局部与全局最优解.我们给出了两个转化条件:一个是充分与必要条件,另一个是在实际中易于验证的充分条件.通过这种转化,我们给出了广义半无限极大极小问题的一个新的一阶最优性条件.
Under certain conditions, we ues the exponential-type augmented Lagrangian function to transform a class of gen- eralized semi-infinite min-max problems into common semi-infinte rain-max problems, with the same set of local and global solutions. We give two conditions for the transformation. One is a necessary and sufficient condition, the other is a sufficient condition which can be verified easily in practice. From the transformation, we obtain a new first-order optimality condition for this class of generalized semi-infinite min-max problems.
出处
《经济数学》
2007年第4期420-426,共7页
Journal of Quantitative Economics
基金
国家自然科学基金资助项目(10571106
10701047)
关键词
标准半无限规划
广义半无限极大极小规划
增广拉格朗日函数
一阶了优性条件
Common semi-infinite programming, generalized semi-infinite rain-max programming, Augmented Lagrangian functions, First-order optimality conditions.