说明：  快速迭代收缩阈值算法，FISTA用一种更为聪明的办法选择序列{xk}，使得其基于梯度下降思想的迭代过程更加快速地趋近问题函数F(x)的最小值。FISTA的缺点是：Lipschitz常数L(f)不一定可知或者可计算。
(Fast iterative contraction threshold algorithm, FISTA uses a more clever method to select the sequence{xk}, making its iterative process based on the gradient descent idea more quickly approach the minimum value of the problem function F(X). The disadvantage of FISTA is that the Lipschitz constant L(f) is not necessarily known or computable.)