摘要
将非线性规划的拟下降方法[3]和无约束优化渐近一维精确搜索规则[4]推广到约束Minimax问题,给出二类拟可行点算法模型及相应的较弱的全局收敛性条件.结果扩充了非线性规划解法的使用范围,为构造求解约束Minimax问题的新的有效方法提供统一的理论依据.
By extending the quasi-descent methods for nonlinear programming and gradually exact line search rule[4] to constrained minimax problems, two algorithm models of quasi-feasible point methods and weak conditions for their global covergence are presented. The results here enlarge the application of nonlinear programming methods and give a unified theory basis for the design of new efficient methods for constrained minimax problem.
出处
《武汉城市建设学院学报》
1996年第1期53-57,共5页
Journal of Wuhan Urban Construction Institute
基金
华中理工大学青年科学基金
关键词
非线性规划
MINIMAX问题
拟可行点法
constrained minimax problem
quasi-feasible point method
global convergence