English

A distance formula for tuples of operators

Functional Analysis 2022-06-06 v1

Abstract

For a tuple of operators A=(A1,,Ad)\boldsymbol{A}= (A_1, \ldots, A_d), dist(A,CdI)\text{dist}(\boldsymbol{A}, \mathbb C^d \boldsymbol{I}) is defined as minzCdAzI\min\limits_{\boldsymbol{z} \in \mathbb C^d} \|\boldsymbol{A-zI}\| and varx(A)\text{var}_x (\boldsymbol{A}) as Ax2j=1dxAjx2.\|\boldsymbol{A} x\|^2-\sum_{j=1}^d {\big|}\langle x| A_j x\rangle{\big|}^2. For a tuple A\boldsymbol{A} of commuting normal operators, it is known that dist(A,CdI)2=supx=1varx(A).\text{dist}(\boldsymbol{A}, \mathbb C^d \boldsymbol{I})^2=\sup_{\|x\|=1}\text{var}_x (\boldsymbol{A}). We give an expression for the maximal joint numerical range of a tuple of doubly commuting matrices. Consequently, we obtain that the above distance formula holds for tuples of doubly commuting matrices. We also discuss some general conditions on the tuples of operators for this formula to hold. As a result, we obtain that it holds for tuples of Toeplitz operators as well.

Cite

@article{arxiv.2206.01503,
  title  = {A distance formula for tuples of operators},
  author = {Priyanka Grover and Sushil Singla},
  journal= {arXiv preprint arXiv:2206.01503},
  year   = {2022}
}

Comments

to appear in LAA

R2 v1 2026-06-24T11:38:08.628Z