摘要
一、引言罚函数方法是数学规划求约束最优解的重要方法之一.自60年代 Zangwill 等人系统地研究罚函数理论以来,发展很快,文献很多.经典的罚函数理论,是通过添加罚函数项后,研究一系列无约束优化问题.并使惩罚参数趋于无限大来获得原规划的最优解.而精确罚函数理论是通过求解单个无约束优化问题来求原规划的最优解.
In this paper,the exact penalty function theory of nonsmooth programming with both e-quality and inequality constraints(?)is discussed.The main results are the following:Suppose X is a compact set,the number of global optimal solution sets of the originalproblem is finite.Under appropriate restrictions,the global exact penalty function exists.Especially,a 1-1 correspondence between the solutions of the original problem and its exactpenalty function problem is set up.Two evaluation theorems for penalty parameters are established.
出处
《系统科学与数学》
CSCD
北大核心
1993年第3期201-210,共10页
Journal of Systems Science and Mathematical Sciences
基金
国家自然科学基金
