Message Board

Dear readers, authors and reviewers,you can add a message on this page. We will reply to you as soon as possible!

2021 Volume 46 Issue 8
Article Contents

LI Yan-yan, JIANG Jian-xin. Error Bounds for Linear Complementarity Problems of ∑1-SDD Matrix and B-∑1-SDD Matrix[J]. Journal of Southwest China Normal University(Natural Science Edition), 2021, 46(8): 1-6. doi: 10.13718/j.cnki.xsxb.2021.08.001
Citation: LI Yan-yan, JIANG Jian-xin. Error Bounds for Linear Complementarity Problems of ∑1-SDD Matrix and B-∑1-SDD Matrix[J]. Journal of Southwest China Normal University(Natural Science Edition), 2021, 46(8): 1-6. doi: 10.13718/j.cnki.xsxb.2021.08.001

Error Bounds for Linear Complementarity Problems of ∑1-SDD Matrix and B-∑1-SDD Matrix

More Information
  • Received Date: 05/03/2020
    Available Online: 20/08/2021
  • MSC: O151.21

  • First, the upper bound of the infinite norm of the inverse matrix ∑1-SDD of A has been studied, Secondly, on the basis of the upper boundary, using the relation ∑1-SDD matrix of A and Ã= I-D+DA, the error bounds for linear complementarity of A is obtained. At the same time, numerical examples are used to illustrate the superiority of the estimator. Lastly, Further obtained the error bounds for linear complementarity of B-∑1-SDD matrix.
  • 加载中
  • [1] PENA J M. A Class of P-Matrices with Applications to the Localization of the Eigenvalues of a Real Matrix[J]. SIAM J Matrix Anal Appl, 2001, 22(4): 1027-1037. doi: 10.1137/S0895479800370342

    CrossRef Google Scholar

    [2] CHEN X J, XIANG S H. Computation of Error Bounds for P-Matrix Linear Complementarity Problems[J]. Math Prog, 2006, 106(3): 513-525. doi: 10.1007/s10107-005-0645-9

    CrossRef Google Scholar

    [3] GAO L, WANG Y Q, LI C Q, et al. Error Bounds for Linear Complementarity Problems of S-Nekrasov Matrices and B-S-Nekrasov Matrices[J]. Journal of Computational and Applied Mathematics, 2018, 336: 147-159. doi: 10.1016/j.cam.2017.12.032

    CrossRef Google Scholar

    [4] DAI P F, LI Y T, LU C J. Error Bounds for Linear Complementarity Problems for SB-Matrices[J]. Numer Algor, 2012, 61(1): 121-139. doi: 10.1007/s11075-012-9533-6

    CrossRef Google Scholar

    [5] GARCÍA-ESNAOLA M, PENA J M. Error Bounds for Linear Complementarity Problem with a ∑-SDD Matrix[J]. Lineaer Algebra and its Application, 2013, 438: 1339-1346. doi: 10.1016/j.laa.2012.09.018

    CrossRef Google Scholar

    [6] LI C Q, LI Y T. Weakly Chained Diagonally Dominant B-Matrices and Error Bounds for Linear Complementarity Problems[J]. Numer Algor, 2016, 73(4): 985-998. doi: 10.1007/s11075-016-0125-8

    CrossRef Google Scholar

    [7] 王峰, 孙德淑. B-矩阵线性互补问题的误差界估计[J]. 数学的实践与认识, 2017, 47(8): 253-260.

    Google Scholar

    [8] 徐玉梅, 王峰. 弱链对角占优B-矩阵线性互补问题的误差界新估计[J]. 西南师范大学学报(自然科学版), 2019, 44(2): 18-24.

    Google Scholar

    [9] 刘毅, 井霞, 高磊. Ostrowski-Brauer Sparse B(OBS-B)矩阵及其线性互补问题的误差界[J]. 云南大学学报(自然科学版), 2021, 43(2): 205-213.

    Google Scholar

    [10] 李艳艳, 周平. Dashnic-Zusmanovich矩阵线性互补问题误差界的估计[J]. 西南师范大学学报(自然科学版), 2019, 44(6): 10-13.

    Google Scholar

    [11] 张小双, 陈震, 刘奇龙. 求解不同阶对称张量组特征值的带位移高阶幂法[J]. 西南大学学报(自然科学版), 2020, 42(8): 81-87.

    Google Scholar

    [12] 文慧敏, 冯德成, 杨亚男. 双参数弱(下)鞅的一类极大值不等式[J]. 西南大学学报(自然科学版), 2021, 43(3): 95-100.

    Google Scholar

    [13] HORN R A, JOHNSON C R. Topics in Matrix Analysis[M]. Cambridge: Cambridge University Press, 1991: 12-36.

    Google Scholar

  • 加载中
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Article Metrics

Article views(3902) PDF downloads(383) Cited by(0)

Access History

Other Articles By Authors

Error Bounds for Linear Complementarity Problems of ∑1-SDD Matrix and B-∑1-SDD Matrix

Abstract: First, the upper bound of the infinite norm of the inverse matrix ∑1-SDD of A has been studied, Secondly, on the basis of the upper boundary, using the relation ∑1-SDD matrix of A and Ã= I-D+DA, the error bounds for linear complementarity of A is obtained. At the same time, numerical examples are used to illustrate the superiority of the estimator. Lastly, Further obtained the error bounds for linear complementarity of B-∑1-SDD matrix.

  • 线性互补问题(Lcp(Aq))的模型是指求$ \boldsymbol{x}\in {{\mathbb{R}}^{n}} $, 满足

    其中A是实矩阵,xq是实向量.

    文献[1]指出:当Lcp(Aq)中的矩阵A是主子式都为正的实矩阵(P矩阵)时,能较容易得到该问题唯一解的误差界.

    文献[2]给出了Lcp(Aq)中的矩阵A是主子式都为正的实矩阵(P矩阵)时的线性互补的误差界

    其中

    关于上述误差界中最难求的$ \mathop {\max }\limits_{\boldsymbol{d} \in {{\left[ {0, 1} \right]}^n}} {\left\| {{{(\boldsymbol{I}-\boldsymbol{D}+\boldsymbol{DA})}^{ - 1}}} \right\|_\infty } $,许多学者进行了卓有成效的研究[3-12].

    本文研究目前少有文献研究的H-矩阵的新子类∑1-SDD矩阵的线性互补问题的误差界估计. 首先给出∑1-SDD矩阵A的逆矩阵无穷范数的上界,其次在该上界的基础上,利用∑1-SDD矩阵AÃ= I - D + DA的关系,得到了A的线性互补问题的误差界.

1.   预备知识
  • $ S\subseteq \mathbb{N}, \bar{S}=\mathbb{N}\backslash S $,对$ \forall i\in \mathbb{N}, {{r}_{i}}(\boldsymbol{A})=\sum\limits_{j\in \mathbb{N}\backslash \left\{ i \right\}}{\left| {{a}_{ij}} \right|}, r_{i}^{S}(\boldsymbol{A})=\sum\limits_{j\in S\backslash \left\{ i \right\}}{\left| {{a}_{ij}} \right|} $.

    文献[5]首次给出了H-矩阵的新子类∑1-SDD矩阵.

    定义1[5]  对于矩阵$ \boldsymbol{A}=\left({{a}_{ij}} \right)\in {{\mathbb{R}}^{n, n}} $,如果存在非空子集$ S\subseteq \mathbb{N} $,使得:

    (ⅰ) $ \left|a_{i i}\right|>r_{i}^{S}(\boldsymbol{A})+1, i \in S $

    (ⅱ) $ \left(\left|a_{i i}\right|-r_{i}^{S}(\boldsymbol{A})-1\right)\left(\left|a_{j j}\right|-r_{j}^{\bar{S}}(\boldsymbol{A})-1\right)>r_{i}^{\bar{S}}(\boldsymbol{A}) r_{j}^{S}(\boldsymbol{A}), i \in S, j \in \bar{S} $.则称A是∑1-SDD矩阵.

    引理1[5]  设A是主对角元素为正的∑1-SDD矩阵,S满足定义1. 则存在对角矩阵W =diag(w1w2,…,wn),且

    $ i\in \mathbb{N}, $γIs,有

    rjS(A)=0时,$ \frac{a_{j j}-r_{j}^{\bar{S}}(\boldsymbol{A})-1}{r_{j}^{S}(\boldsymbol{A})}=\infty $. 则AW是严格对角占优矩阵.

    引理2[13]  设矩阵A是H-矩阵,则|A-1|≤〈A-1. 其中

    AB指的是aijbij$ i, j\in \mathbb{N} $.

2.   ∑1-SDD矩阵无穷范数的上界估计
  • 本部分利用构造的方法,给出仅与矩阵元素有关的∑1-SDD矩阵无穷范数的上界.

    定理1  设矩阵$ \boldsymbol{A}=\left( {{a}_{ij}} \right)\in {{\mathbb{R}}^{n, n}} $是∑1-SDD矩阵,则

      由引理2知|A-1|≤〈A-1,则

    则〈Ax=e=(1,…,1)T,定义$ {x_{{i_0}}} = \mathop {\max }\limits_{i \in S} {x_i}, {x_{{j_0}}} = \mathop {\max }\limits_{j \in \bar S} {x_j} $,由〈Ax=e

    对于

    同理得

    联合(1),(2)式,整理化简得

3.   ∑1-SDD矩阵线性互补问题的误差界
  • 本部分利用定理1中∑1-SDD矩阵逆矩阵无穷范数的上界,得到它线性互补问题的上界估计式.

    定理2  设矩阵$ \boldsymbol{A}=\left( {{a}_{ij}} \right)\in {{\mathbb{R}}^{n, n}} $是∑1-SDD矩阵,$ S\subseteq \mathbb{N}, S\ne \mathit{\varnothing} , {{a}_{ii}}>0 $,令$ \boldsymbol{\tilde{A}}=\boldsymbol{I}-\boldsymbol{D}+\boldsymbol{DA}=\left( {{{\tilde{a}}}_{ij}} \right), \boldsymbol{D}=\text{diag}\left( {{d}_{i}} \right), 0 <{{d}_{i}}\le 1 $,则对$ \forall i, j\in \mathbb{N} $,有

    且对$ i \in S, j \in \bar S $,有

    即矩阵Ã是∑1-SDD矩阵.

      由Ã的定义知,(4)式显然成立.

    下面重点证明(5)式. 由(4)式和Ã的定义知

    又由∑1-SDD矩阵的定义知aii>riS(A)+1,所以$ \left| {{{\tilde{a}}}_{ii}} \right|>r_{i}^{S}\left( {\boldsymbol{\tilde{A}}} \right)+1 $.

    同理对任意jS,有

    则由定义1知,Ã是∑1-SDD矩阵.

    定理3  设$ \boldsymbol{A}=\left( {{a}_{ij}} \right)\in {{\mathbb{R}}^{n, n}} $是∑1-SDD矩阵,$ S\subseteq \mathbb{N}, S\ne \mathit{\varnothing} , {{a}_{ii}}>0 $,令

      令Ã= I - D + DA =(ãij),由的定义知,对任意iSjS,有

    则对任意iSjS,有

    同理对iSjS,类似地有

    则由定理1知

    例1  设$ {{\boldsymbol{A}}_{1}}=\left( \begin{matrix} 16 & -9 & -9 \\ -9 & 27 & -9 \\ -9 & -9 & 27 \\ \end{matrix} \right) $. 经验证,当S={1}时,A1为∑1-SDD矩阵,且$ {{I}_{s}}=\left[ \frac{6}{5}, \frac{17}{9} \right) $. 应用定理3,当didj=1时,计算得

    例2  设$ {{\boldsymbol{A}}_{2}}=\left( \begin{matrix} 4 & -1 & 0 & -1 \\ -1 & 6 & -1 & 0 \\ -2 & -1 & 8 & -1 \\ \frac{1}{2} & \frac{1}{2} & \frac{1}{4} & 4 \\ \end{matrix} \right) $,经验证,当S={1,2}时,A2为∑1-SDD矩阵,且$ {{I}_{s}}=\left[ \frac{1}{2}, 2 \right) $. 应用定理3,当didj=1时,计算得

    本部分给出了形式比较简洁,易于计算的线性互补误差界的上界.

4.   B-∑1-SDD矩阵误差界的估计
  • $ \boldsymbol{A}=\left( {{a}_{ij}} \right)\in {{\mathbb{R}}^{n, n}}, \boldsymbol{A}={{\boldsymbol{B}}^{+}}+\boldsymbol{C} $

    B+是Z-矩阵(非主对角元素非正的矩阵),C是非负矩阵.

    定义2  若矩阵$ \boldsymbol{A}=\left( {{a}_{ij}} \right)\in {{\mathbb{R}}^{n, n}} $能写成(6)式的形式,且B+是主对角元素为正的∑1-SDD矩阵,则称A是B-∑1-SDD矩阵.

    引理3  若A是B-∑1-SDD矩阵,则A是P-矩阵.

    定理4  设$ \boldsymbol{A}=\left( {{a}_{ij}} \right)\in {{\mathbb{R}}^{n, n}} $是B-∑1-SDD矩阵,$ \mathit{\varnothing} \ne S\subset \mathbb{N}, {{\boldsymbol{B}}^{+}} $如(6)式定义,则

      因为A是B-∑1-SDD矩阵,A = B++CB+是主对角元素为正的∑1-SDD-Z矩阵. 设矩阵D =diag(di),0 < di≤1,

    由文献[11]知$ \boldsymbol{\tilde{B}}_{D}^{+} $仍是主对角元素为正的∑1-SDD矩阵,且$ \boldsymbol{\tilde{B}}_{D}^{+} $是非奇异M-矩阵. 由文献[11]中定理2的证明知

    因为$ \boldsymbol{\tilde{B}}_{D}^{+} $是∑1-SDD矩阵,由定理1的证明知

    本文首先给出了∑1-SDD矩阵逆的无穷范数的新上界,其次研究了目前较少有文献研究的∑1-SDD矩阵和B-∑1-SDD矩阵的线性互补问题的误差界,这是对该类矩阵这方面研究空白的填充,也是对线性互补误差界问题研究的进一步拓展.

Reference (13)

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return