Let ${\cal V}$ be n Hilbert space over $\mathbb{R}$ or $\mathbb{C}$, with an orthonormal basis $(e_n)_{n=1}^{\infty}$. For every bounded linear operator $A: {\cal V} \to {\cal V}$, we can associate $A$ with an matrix $(a_{ij})_{i,j=1}^{\infty}$ where $a_{ij} = \langle Ae_j,e_i\rangle$.
Is there a matrix $(b_{ij})_{i,j=1}^{\infty}$ such that:
1. $\sum_{i=1}^{\infty}|b_{ij}|^2 < \infty$
2. $\sum_{j=1}^{\infty}|b_{ij}|^2 < \infty$
3. $\sup_{i,j \ge 1} |b_{ij}| < \infty$
and there is no bounded linear operator $B : {\cal V} \to {\cal V}$ such that $b_{ij} = \langle Be_j,e_i\rangle$
I am really struggling with this problem, I will appreciate any help!
Try this matrix... $$ B = \begin{bmatrix} 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 &\cdots\\ 0 & 1 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0\\ 0 & 0 & 0 & 1 & 1 & 1 & 0 & 0 & 0 & 0 & 0\\ 0 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 1 & 0\\ \vdots \end{bmatrix} $$ Each column has a single $1$ and all the rest zeros.
The first row has one $1$ and all the rest zeros.
The second row has two $1$s and all the rest zeros.
The third row has three $1$s and all the rest zeros.
The fourth row has four $1$s and all the rest zeros.
...
The $k$th row has $k$ $1$s and all the rest zeros.
The three conditions are satisfied.
Think of using Hilbert space $l_2$, with its canonical orthonormal basis $(e_k)$. Think of the elements as infinite column vectors.
Fix $k$. Let vector $x$ be the transpose of row $k$. Then $Bx = k e_k$. But $\|x\| = k^{1/2}$ and $\|Bx\| = k$, so for the operator norm of $B$ we conclude $$ \|B\| \ge \frac{k}{k^{1/2}} = k^{1/2}. $$ This holds for all $k$. Thus $B$ is not a bounded operator.