错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Estimates of Euclidean numerical radius for block matrices

  • Pintu Bhunia,
  • Suvendu Jana,
  • Kallol Paul

摘要

We develop several Euclidean numerical radius bounds for the product of two d-tuple operators using positivity criteria of a \(2\times 2\) 2 × 2 block matrix whose entries are d-tuple operators. From these bounds, by using polar decomposition of operators, we obtain Euclidean numerical radius bounds for d-tuple operators. Among many other bounds, it is shown that \(\begin{aligned} w_e(\textbf{A})\le & {} \frac{1}{\sqrt{2}} \Vert \textbf{A}\Vert ^{1/2}\sqrt{\left\| \sum _{k=1}^{d} (|A_k|+|A_k^*|)\right\| }, \end{aligned}\) w e ( A ) 1 2 A 1 / 2 k = 1 d ( | A k | + | A k | ) , where \(w_e(\textbf{A})\) w e ( A ) and \(\Vert \textbf{A}\Vert \) A are the Euclidean numerical radius and the Euclidean operator norm, respectively, of a d-tuple operator \(\textbf{A}=(A_1,A_2, \ldots ,A_d).\) A = ( A 1 , A 2 , , A d ) . Further, we develop an upper bound for the Euclidean numerical radius of an \(n\times n\) n × n operator matrix whose entries are d-tuple operators. In particular, it is proved that if \([\mathbf {A_{ij}}]_{n\times n}\) [ A ij ] n × n is an \(n\times n\) n × n operator matrix then \(\begin{aligned} w_e([\mathbf {A_{ij}}]_{n\times n})\le w ([a_{ij}]_{n\times n}), \end{aligned}\) w e ( [ A ij ] n × n ) w ( [ a ij ] n × n ) , where each \(\mathbf {A_{ij}}\) A ij is a d-tuple operator, \(1\le i,j\le n\) 1 i , j n , \(a_{ij}=w_e(\mathbf {A_{ij}})\) a ij = w e ( A ij ) if \(i=j\) i = j , \(a_{ij}= \sqrt{w_e(|\mathbf {A_{ji}|}+|\mathbf {A_{ij}^*}|)w_e(|\mathbf {A_{ij}|}+|\mathbf {A_{ji}^*}|)}\) a ij = w e ( | A ji | + | A ij | ) w e ( | A ij | + | A ji | ) if \(i<j\) i < j , and \(a_{ij}= 0\) a ij = 0 if \(i>j\) i > j . Some related applications are also discussed.