摘要
对线性或非线性约束的非线性规划问题给出了一个简便的既约梯度法,并在一定的条件下,证明了此算法的全局收敛性。
A simple reduced gradient algorithm for nonlinear programming with linearor nonlinear constraints is given.Under the assumption that all feasible points but K-Tpoints satisfy the regularity condition,It is proved that the algorithm either terminates at aK-T point with finite steps orgenerates aninfinites sequence of feasible points whose everycluster point is a K-T point.
出处
《曲阜师范大学学报(自然科学版)》
CAS
1994年第4期1-6,共6页
Journal of Qufu Normal University(Natural Science)
基金
国家自然科学基金
关键词
最优化
既约梯度法
正则性
非线性规划
optimization K-T point regularity condition reduced gradientalgorithm