Parallel and cyclic hybrid subgradient extragradient methods for variational inequalities
Parallel and cyclic hybrid subgradient extragradient methods for variational inequalities
复制标题
变分不等式的并行和循环混合次梯度超梯度方法
DOI：
10.1007/s13370-016-0473-5
复制
发表时间：
2016
影响因子：
1.1
通讯作者：
D. Hieu
中科院分区：
文献类型：
--
作者：
D. Hieu
本文提出了两种求解变分不等式公共解的混合次梯度外梯度方法。第一种是并行算法，可以同时进行，而第二种是循环算法，它是按顺序计算的家庭中的每个子问题。本文的新奇在于，我们设计的算法，以开发可能的实用数值方法时，子问题的数量是大的。这些算法可以被认为是对CSVIPs一些已知结果的改进。数值实验也说明了所提出的算法的效率。
In this paper, we propose two hybrid subgradient extragradient methods for solving common solutions of variational inequalities problems (CSVIP). The first is a parallel algorithm which can be performed simultaneously while the second is a cyclic algorithm which is computed sequentially on each subproblem in the family. The novelty of this paper is that we have designed the algorithms to develop possible practical numerical methods when the number of subproblems is large. The algorithms can be considered as improvements of some previously known results for CSVIPs. Numerical experiments are also performed to illustrate the efficiency of the proposed algorithms.