摘要
本文利用开关函数．建立了解线性约束优化问题的一个组合型可行方向法─—开关算法模型，并给出了其收敛性质，从而统一、推广了包括起线性收敛的算法在内的常见的可行方向法．依此模型，具体构造了一类起线性收敛的新算法．
By using the concept of switch function, we propose a hybrid methed of feasibledirections──a model of switch algorithms for linearly constrained optimization andgive its convergence. The medel unifies and extends the common feasible directionmethods including ones with superlinear convergence. We specially give a class of newalgorithms with superlinear convergence, in which two switch funtions are used.
出处
《应用数学》
CSCD
1997年第2期14-18,共5页
Mathematica Applicata
关键词
线性约束化
最佳化
开关算法模型
收敛性
Linearly constrained optimization, Feasible direction methed
Unified approach
Model of switch algorithms
Convergence
