非线性不等式约束优化问题的指数型精确罚函数算法  

Algorithm of exponential-type exact penalty function for nonlinear inequality constrained optimization

作　　者：杨莲 姚奕荣[1] YANG Lian;YAO Yirong(College of Sciences, Shanghai University, Shanghai 200444, Chin)

机构地区：[1]上海大学理学院,上海200444

出　　处：《上海大学学报（自然科学版）》2017年第6期911-919,共9页Journal of Shanghai University:Natural Science Edition

摘　　要：针对非线性不等式约束优化问题,通过增加一个变量构造了一种新的指数型罚函数,进而证明了该罚函数的光滑性和精确性.进一步,设计了一种求解非线性不等式约束优化问题的精确罚函数算法.数值计算的结果表明了该算法的可行性.To deal with nonlinear inequality constrained optimization problems, we con- struct a new exponential-type penalty function by adding a variable, and prove smoothness and accuracy of the penalty function. In addition, an exact penalty function algorithm is proposed to solve nonlinear inequality constrained optimization problems. Numerical re- sults are reported to show effectiveness of the algorithm.

关 键 词：非线性优化 光滑精确罚函数 罚函数算法 M-F(Mangasarian-Fromovitz)条件 

分 类 号：O221.2[理学—运筹学与控制论]