摘要
本文提出求解不可微非线性不等式约束极小化问题的 L_1-精确罚函数算法。在有关函数为半光滑的假设下,给出了收敛性结果。
A new kind of algorithm for minimizing an objective function subject to inequality constraints is presented,the functions relative to the problem are Lipschitz continuous.The algorithm determines a search direction by solving a quadratic programming subproblem,which has always an optimal solution,and uses an exact penalty function to compute the steplength along this direction through an Armijo-type scheme.Convergence results have been presented under the condition that all of the functions relative to the problem are semismooth.
出处
《宁波大学学报（理工版）》
CAS
1992年第1期1-8,共8页
Journal of Ningbo University：Natural Science and Engineering Edition
关键词
不可微
非线性规划
罚函数法
收敛性
Nondifferentiable
constrained minimization
exact penalty functions
sequential quadratic programming
covergence
