Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

If have the following problem:

Let $A : \mathcal{H} \to \mathcal{H}$ be a bounded, self-adjoint operator on some Hilbert space $\mathcal{H}$. Let $B: \mathcal{H} \to \mathcal{H}$ be a bounded, strictly positive operator $B \geq c > 0$. Is the following statement true or false: $$ 0 \in \sigma_{ac}(A) \Leftrightarrow 0 \in \sigma_{ac}(BAB) \quad ? $$ Or can one say something about the relation of $\lim_{y \searrow 0} \Im \langle f, (A- i y)^{-1} f \rangle$ and $\lim_{y \searrow 0} \Im \langle f, (A- iB y)^{-1} f \rangle$. $A$ and $B$ do not commute.

I'm aware of a similar question: Spectrum of the operator PAP, with A self-adjoint and P strictly positive. I hope that the one I state here is a little more specific as to come up with a counterexample or a proof.

share|improve this question
1  
I do not know if it helps, but at least in this case you can argue that $BAB$ is similar to $B^2A$ or $AB^2$. –  András Bátkai Aug 12 '13 at 10:38
add comment

1 Answer

This is false. Let $A=S+S^*$ be the free Jacobi matrix on $\ell^2$; I write $S$ for the shift $(Sy)_n = y_{n+1}$. Then $\sigma_{ac}(A)=[-2,2]$. Let $B$ be multiplication by a bounded sequence $b_n\ge c$. Then $$ BAB = b(S+S^*)b = bb_+S + bb_-S^* $$ is another Jacobi matrix, and pretty much any choice of $b$ provides a counterexample as it's so easy to destroy ac spectrum of one-dimensional operators. For example, any $b$ that takes only two values and is not eventually periodic works: $\sigma_{ac}(BAB)=\emptyset$

See perhaps my paper http://annals.math.princeton.edu/2011/174-1/p04 for background on this.

share|improve this answer
    
Is this related to anderson localisation ? –  jjcale Apr 1 at 19:03
    
Not very directly the way I did it, but I could also finish the argument by referring to Anderson localization (pick a random $b$). –  Christian Remling Apr 1 at 19:09
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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