共轭梯度法（Conjugate Gradient）是介于最速下降法与牛顿法之间的一个方法，它仅需利用一阶导数信息，但克服了最速下降法收敛慢的缺点，又避免了牛顿法需要存储和计算Hesse矩阵并求逆的缺点，共轭梯度法不仅是解决大型线性方程组最有用的方法之一，也是解大型非线性最优化最有效的算法之一。这里给出共轭梯度法的源程序-Conjugate gradient method (Conjugate Gradient) is between the steepest descent method and Newton' s method between a method that takes only a first derivative information, but to overcome the slow convergence of the steepest descent method shortcomings, but also avoid the need to store Newton and computing the inverse Hesse matrix and disadvantages, conjugate gradient method is not only to solve large linear equations of the most useful methods, large-scale nonlinear optimization solution is the most efficient algorithm. Here is the source conjugate gradient method