发表机构
Xiangtan University; Huaibei Normal University(湘潭大学; 淮北师范大学)
机构由 AI 辅助整理,请以论文原文为准。AI 中文总结
构造两个3×3正对称矩阵,利用Sturm定理证明黄氏弱受控猜想在k=2时失效,给出反例。
AI 中文摘要
我们给出了Z. Huang(线性代数及其应用,434卷,第2期,2011年,第457-462页)提出的关于非负矩阵Hadamard积奇异值的弱受控问题的一个反例。该问题询问:对于所有非负矩阵$A$和$B$,不等式$\bigl\{s_j^2(A\circ B)\bigr\} \prec_w \bigl\{s_j(A\circ A)s_j(B\circ B)\bigr\}$是否成立,其中$\circ$表示Hadamard积,$\prec_w$表示弱受控,$s_j(\cdot)$表示矩阵的第$j$大奇异值。为回答此问题,我们构造了两个$3\times3$逐项正对称矩阵。我们利用Sturm定理给出了相关奇异值和的严格上下界,并证明当$k=2$时该不等式对部分和失效。
英文摘要
We give a counterexample to a weak majorization problem, proposed by Z. Huang (Linear Algebra Appl., 434 (2) (2011) 457--462), for singular values of Hadamard products of nonnegative matrices. The problem asks whether, for all nonnegative matrices $A$ and $B$, \begin{equation*} \bigl\{s_j^2(A\Had B)\bigr\} \prec_w \bigl\{s_j(A\Had A)s_j(B\Had B)\bigr\} \end{equation*} holds, where $\Had$ denotes the Hadamard product, $\prec_w$ means weak majorization, and $s_j(\cdot)$ is the $j$th largest singular value of a matrix. To answer this problem, we construct two $3\times3$ entrywise positive symmetric matrices. We give rigorous upper and lower bounds for the relevant singular-value sums by Sturm's theorem, and show that the inequality fails for the partial sum with $k=2$.
Comments5 pages