CNJK鞍点共轭梯度法

Conjugate Gradient Method of CNJK Saddle Point

提出一种新的共轭梯度法，它可在有限步内求得二次函数Ｚ（Ｘ，λ）＝λ￣（Ｔ）ＡＸ＋Ｃ￣（Ｔ）Ｘ＋ｂ￣（Ｔ）λ的鞍点值。

The authors gave a new algorithm conjugate gradient method of saddle point. It could be used to solve the quadratic function Z(X,λ)=λ￣(T)AX+C￣(T)X+b￣(T)λ.The saddle point value of the function may be obtained using the method with finite iterative steps.