一种解决大系统奇异控制问题的新算法
A New Algorithm for Solving Singular Control Problems in Large Scale Systems
-
摘要: 本文提出了一个解决奇异控制问题的新算法——诱导梯度法.本文证明了这个算法的收敛性.共轭梯度法适用于求解函数Z=ax2+by2的极值点问题,本文提出的算法适用于求解函数Z=axy+bx+cy+d的鞍点问题.Abstract: In this paper a new algorithm-Leading Directional Derivative gradient method with its convergency proved theoretically is presented that is formulated to solve singular control problems.The method presented here is applicable to calculating saddle points of the function Z=axy+bx+cy+d,while the conjugate gradient method is applicable to calculating extreme values of the function Z=ax2+by2.