7
$\begingroup$

Suppose we have a separable Hilbert space (thus with a countable basis) and that represent an operator in matrix form, i.e:

$A: H \rightarrow H $$$x \;\rightarrow \sum_{j \in \mathbb{N}}\left(\sum_{k \in \mathbb{N}} a(j,k)\cdot\langle x,e_k\rangle \right)e_j$$

Given that:

  • The series of complex numbers $\sum_{k \in \mathbb{N}} a(j,k)\cdot\langle x,e_k\rangle $ converges.

  • That $\sum_{j \in \mathbb{N}}\left(\sum_{k \in \mathbb{N}} a(j,k)\cdot\langle x,e_k\rangle \right)e_j$ converges.

  • That $\sum_{j \in \mathbb{N}}\sum_{k \in \mathbb{N}} \left |a(j,k) \right|^2 < \infty$ (the previous conditions + this one imply that A is a Hilbert-Schmidt operator).

  • And that $\sum_{k \in \mathbb{N}} \left |a(k,k) \right| < \infty$.

Prove that $A$ is a trace class operator.

My attempt at a solution

By these conditions we know that $\left | Tr (A) \right | \leq \displaystyle{\sum_{k \in \mathbb{N}}} \left | a(k,k) \right | < \infty$. But I can't see the connection to $Tr(\left |A \right |)$. Any hints on how to proceed?

$\endgroup$
8
$\begingroup$

I don't think that $A$ needs to be trace-class.

Consider a sequence $b=\{b_j\}$ such $b\in \ell^2(\mathbb N)\setminus \ell^1(\mathbb N)$ (for example, $b_j=1/j$). Define the operator $$ A=\bigoplus_j\begin{bmatrix}0&b_j\\ \overline {b_j}&0\end{bmatrix}. $$ That is, in the canonical basis, $$ A=\begin{bmatrix}0&1\\ 1&0 \\& & 0 & 1/2\\ & & 1/2&0\\ &&&&\ddots \end{bmatrix} $$ where the empty entries are also $0$.

Since every row contains a single nonzero element, the first condition is satisfied trivially. The second condition is also easy, because we end up with a sum over the canonical basis with coefficients in an $\ell^2$-sequence.

Third condition holds, since $\sum_j\sum_k|a(j,k)|^2=2\sum_j|b_j|^2<\infty$.

And the fourth condition holds because $a(k,k)=0$ for all $k$.

But $A$ is not trace-class: because $|A|$ is the diagonal operator with diagonal $b_1,b_1,b_2,b_2\ldots$, and so its trace is infinite.

$\endgroup$
  • 2
    $\begingroup$ This seems to be a canonical counterexample, +1. $\endgroup$ – Julien Apr 19 '13 at 2:10
  • 1
    $\begingroup$ I guess. I actually learnt it from an answer to a question of mine: math.stackexchange.com/questions/174821/… $\endgroup$ – Martin Argerami Apr 19 '13 at 5:30
  • $\begingroup$ Nice counterexample. Thanks! $\endgroup$ – grizzlyjoker Apr 19 '13 at 7:35
  • 1
    $\begingroup$ @MartinArgerami What does the notation $|A|$ mean here? $\endgroup$ – ManUtdBloke Jul 16 '17 at 11:43
  • 2
    $\begingroup$ @eurocoder: $|A|=(A^*A)^{1/2}$. $\endgroup$ – Martin Argerami Jul 16 '17 at 11:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.