H-矩阵非线性互补问题基于模的矩阵分裂迭代法改进的收敛性定理
The Improved Convergence Theorems of Modulus-Based Matrix Splitting Iteration Methods for a Class of Nonlinear Complementarity Problems with H-Matrices
通讯作者:
收稿日期: 2021-01-29
基金资助: |
|
Received: 2021-01-29
Fund supported: |
|
In this paper, we proved the convergence theories of the modulus-based matrix splitting iteration methods and the corresponding acceleration method for nonlinear complementarity problems of
Keywords:
本文引用格式
马昌凤, 马飞洋.
Ma Changfeng, Ma Feiyang.
1 引言
其中
显然, 当
由于对角可微函数
本文通过给出一个较弱的条件, 证明了
本文的其余部分组织如下. 第2节给出了一些有用的记号、定义和引理. 第3节, 我们对基于模的矩阵分裂迭代法进行了新的收敛性分析. 第4节给出了一些算例. 第5节是小结.
2 预备知识
本节先简要回顾几个必要的符号和引理.
对于两个矩阵
对于
方阵
对于非奇异矩阵
下面的引理给出了
引理2.1[29] 设
3 算法及改进的收敛性定理
在本节中, 我们将给出新的收敛性分析, 以改进NCP (1.1) 中当
为了进行下面的讨论, 我们假设
令
和
取
其中
算法3.1 设
这里
在文献[31] 中, Zheng针对
定理3.1 设
证 设
要证
利用(3.5) 式和
令
不难发现,
注3.1 注意到
由
不是
在算法3.1中, 每次迭代都需要求解线性方程组(3.3). 在实际执行过程中要得到精确解几乎是不现实的, 特别是对于大型稀疏线性方程组. 因此, 用加速迭代法近似求解(3.3) 是比较理想的选择. 注意到, 若
其中
算法3.2 设
并设
在下文中, 我们给出了当系统矩阵
定理3.2 设
(a) 当
(b) 当
证 (a) 首先证明
根据
结合(3.8) 和(3.9) 式可得
由于
将(3.11) 式代入(3.10) 式, 我们有
经过简单的计算, (3.12) 式是可重写为
因为
令
因
(b) 类似地, 我们证明
由于
结合(3.15) 和(3.10)式, 我们有
注意到(3.16) 式可以重写为
由前面的分析, 立即可得
令
因此, 若
注3.2 虽然在条件
注3.3 对于文献[32] 中基于模的矩阵分裂加速迭代法, 当矩阵
4 数值实验
本节将提供一些数值例子来验证新定理的理论分析. 实验结果中, "Iter"表示迭代步数, "CPU"表示所耗费的计算时间, "Res"表示残差向量的模, 即
其中min算子是按分量取极小. 在所有的算例中, 随机选取初始向量
4.1 算法3.1的数值实验
例4.1 求向量
其中
非线性函数
取
实验结果见表 1.
表 1 例4.1的实验结果
非线性函数 | Iter | CPU | Res |
26 | 0.0010 | 8.5203e-006 | |
26 | 0.0010 | 8.0783e-006 | |
27 | 0.0010 | 7.0378e-006 | |
24 | 0.0001 | 8.9217e-006 | |
27 | 0.0370 | 6.7821e-006 |
4.4 算法3.2的数值实验
例4.2 求
其中
令
利用算法3.2, 迭代17次即可获得原问题满足终止准则的近似解. 此外, 不难发现,
例4.3 问题的矩阵
考虑
不难发现, 这两个分裂都是
表 2 例4.1的实验结果
情形1: | 情形2: | |||||
Iter | CPU | Res | Iter | CPU | Res | |
26 | 0.0010 | 8.5203e-006 | 26 | 0.0010 | 8.5203e-006 | |
26 | 0.0010 | 8.0783e-006 | 26 | 0.0010 | 8.5203e-006 | |
27 | 0.0010 | 7.0378e-006 | 26 | 0.0010 | 8.5203e-006 | |
24 | 0.0001 | 8.9217e-006 | 26 | 0.0010 | 8.5203e-006 | |
27 | 0.0370 | 6.7821e-006 | 26 | 0.0010 | 8.5203e-006 |
5 小结
定理3.1给出了在较弱条件下求解非线性互补问题的基于模的矩阵分裂迭代方法的收敛理论. 由于
同样, 定理3.2也改进了基于加速模的矩阵分裂迭代法在新条件下的收敛性理论. 这表明对分裂
参考文献
Engineering and economic applications of complementarity problems
DOI:10.1137/S0036144595285963 [本文引用: 1]
Parallel nonlinear AOR method and its convergence
DOI:10.1016/0898-1221(95)00190-5 [本文引用: 1]
On smoothing methods for the
DOI:10.1137/S1052623498335080 [本文引用: 1]
Modulus-based matrix splitting methods for linear complementarity problems
Two-step modulus-based matrix splitting iteration method for linear complementarity problems
DOI:10.1007/s11075-010-9416-7 [本文引用: 2]
A general modulus-based matrix splitting method for linear complementarity problems of
DOI:10.1016/j.aml.2013.06.015 [本文引用: 1]
Accelerated modulus-based matrix splitting iteration methods for linear complementarity problem
DOI:10.1007/s11075-012-9664-9 [本文引用: 1]
Two-stage multisplitting iteration methods using modulus-based matrix splitting as inner iteration for linear complementarity problems
DOI:10.1007/s10957-013-0362-0 [本文引用: 1]
On the convergence analysis of two-step modulus-based matrix splitting iteration method for linear complementarity problems
A relaxation modulus-based matrix splitting iteration method for solving linear complementarity problems
On the monotone convergence of matrix multisplitting relaxation methods for the linear complementarity problem
Matrix multisplitting methods with applications to linear complementarity problems: Parallel synchronous and chaotic methods
Matrix multisplitting methods with applications to linear complementarity problems: Parallel asynchronous methods
Modulus-based synchronous multisplitting iteration methods for linear complementarity problems
DOI:10.1002/nla.1835
Modulus-based synchronous two-stage multisplitting iteration methods for linear complementarity problems
A modified modulus method for symmetric positive-definite linear complementarity problems
DOI:10.1002/nla.609
Nonstationary extrapolated modulus algorithms for the solution of the linear complementarity problem
A preconditioned modulus-based iteration method for solving linear complementarity problems of
A general accelerated modulus-based matrix splitting iteration method for solving linear complementarity problems
Improved convergence theorems of modulus-based matrix splitting iteration methods for linear complementarity problems
The modulus-based nonsmooth Newton's method for solving linear complementarity problems
DOI:10.1016/j.cam.2015.04.006 [本文引用: 1]
Engineering and economic applications of complementarity problems
DOI:10.1137/S0036144595285963 [本文引用: 1]
Finite-dimensional variational inequality and nonlinear complementarity problems: a survey of theory, algorithms and applications
DOI:10.1007/BF01582255 [本文引用: 1]
Modulus-based matrix splitting iteration methods for a class of nonlinear complementarity problem
Two-step modulus-based matrix splitting iteration method for a class of nonlinear complementarity problems
DOI:10.1016/j.laa.2016.01.002 [本文引用: 1]
Modified modulus-based matrix splitting algorithms for a class of weakly nondifferentiable nonlinear complementarity problems
DOI:10.1016/j.apnum.2016.05.004 [本文引用: 3]
Accelerated modulus-based matrix splitting iteration methods for a restricted class of nonlinear complementarity problems
DOI:10.1007/s11075-016-0243-3 [本文引用: 1]
Improved convergence theorems of modulus-based matrix splitting iteration method for nonlinear complementarity problems of
DOI:10.1007/s10092-017-0236-1 [本文引用: 3]
Accelerated modulus-based matrix splitting iteration method for a class of nonlinear complementarity problems
DOI:10.1007/s40314-017-0496-z [本文引用: 4]
/
〈 |
|
〉 |
