Conjugate-gradient

  共轭梯度法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。(Conjugate gradient method is between the steepest descent method and Newton method between a method that only use the first derivative information, but the steepest descent method to overcome the disadvantage of slow convergence, but also avoids the need to store and calculate Newton Hesse matrix and the shortcomings of the inverse, conjugate gradient method is not only linear equations to solve large-scale one of the most useful, large-scale nonlinear optimization solution is also the most efficient algorithms.)