求解极大单调包含问题的改进混合外梯度邻点方法

A Modified Hybrid Proximal-Extragradient Method for Maximal Monotone Inclusions

提出一种改进的求解极大单调包含问题的混合外梯度邻点方法.该方法的每个迭代步都通过Armijo准则得到正则化参数,在不要求其下界严格大于零的条件下,证明了该方法的弱收敛性,并给出了改进方法的两个特例.

We proposed a modified hybrid approximate extragradient-proximal point algorithm. At each iteration, the regularization parameter, obtained by means of Armijo-like condition, is not necessarily bounded away from zero, while the method＇s weak global convergence still exists. Finally, we gave two special cases of the modified framework.