Gaussian-measurement equality with the Holevo bound

Unsolved ID op_8f1853475db7ea27 Last edited 10 September 2026
Edit

Problem

When does the optimal single-copy Gaussian-measurement cost equal the Holevo bound? Give necessary and sufficient conditions in terms of the local first moments, covariance, their first derivatives, and the weight matrix defined below. Let \(\rho_\theta\) be a smooth faithful \(n\)-mode Gaussian model, with integers \(n,p\geq1\) and \(\theta\in\Theta\subset\mathbb R^p\) for an open set \(\Theta\). Write \(R=(q_1,p_1,\ldots,q_n,p_n)^{\mathsf T}\) for the canonical quadrature vector. Its mean vector and covariance are defined in Eq. (1). The commutation convention is Eq. (2).

\begin{equation} \begin{aligned} (d_\theta)_j&:=\operatorname{Tr}(\rho_\theta R_j),\\ (V_\theta)_{jk}&:=\frac12\operatorname{Tr}(\rho_\theta\{R_j-(d_\theta)_j,R_k-(d_\theta)_k\}). \end{aligned} \tag{1} \end{equation}
\begin{equation} [R_j,R_k]=i\Omega_{jk},\qquad \Omega:=\bigoplus_{j=1}^{n}\begin{pmatrix}0&1\\-1&0\end{pmatrix}. \tag{2} \end{equation}

Assume the symmetric logarithmic derivative quantum Fisher information matrix is nonsingular. Let \(W\) be a real positive-definite \(p\times p\) weight matrix.

Define the optimal single-copy Gaussian-measurement cost by Eq. (3).

\begin{equation} C_G(\theta,W):=\inf_{M\ \mathrm{Gaussian}}\operatorname{tr}(WF_M(\theta)^{-1}), \tag{3} \end{equation}

where \(F_M\) is the classical Fisher information matrix of measurement \(M\), and singular \(F_M\) has infinite cost. Gaussian measurements mean Gaussian-ancilla preparations followed by Gaussian unitaries and homodyne detection with arbitrary classical processing.

Define the Holevo bound \(C_H(\theta,W)\) as the infimum of Eq. (4).

\begin{equation} \operatorname{tr}(W\operatorname{Re}Z)+\|\sqrt W\operatorname{Im}Z\sqrt W\|_1 \tag{4} \end{equation}

The infimum is over Hermitian observables \(X_1,\ldots,X_p\) with finite second moments, subject to Eq. (5).

\begin{equation} \begin{aligned} Z_{jk}&:=\operatorname{Tr}(\rho_\theta X_jX_k),\\ \operatorname{Tr}(\rho_\theta X_j)&=0,\qquad \operatorname{Tr}((\partial_k\rho_\theta)X_j)=\delta_{jk}. \end{aligned} \tag{5} \end{equation}

The measurement is optimized locally at the specified \(\theta\); its setting is held fixed when taking derivatives for \(F_M\). The symmetric logarithmic derivatives \(L_j\) satisfy \(2\partial_j\rho_\theta=\rho_\theta L_j+L_j\rho_\theta\). Their Fisher matrix has entries \(\operatorname{Re}\operatorname{Tr}(\rho_\theta L_jL_k)\). The requested criterion must characterize \(C_G(\theta,W)=C_H(\theta,W)\), rather than merely restate these two optimizations. Equality of the infima is the target. A limiting sequence of Gaussian measurements counts, even if no individual measurement attains the value.

Source

This attainability classification is derived from the Gaussian-measurement restriction identified as an open direction in the Discussion of Chang, Genoni, and Albarelli [CGA26]. It is narrower than asking how to evaluate the Holevo bound itself.

Progress

  • The Holevo minimization for a faithful Gaussian model can be restricted exactly to centered linear and symmetrized quadratic quadrature observables. Chang, Genoni, and Albarelli prove the required invariant-subspace property in Derivations and give a finite semidefinite program in Eqs. (36) and (72) [CGA26]. This computes \(C_H\) from the local moments and derivatives. It does not show that a Gaussian measurement attains it.

  • For Gaussian shift models with parameter-independent covariance, Gaussian measurements attain the Holevo bound. Bradshaw, Lam, and Assad construct the measurement in Sec. III.B and extend it to arbitrary finite mode and displacement-parameter counts in Appendix C [BLA18]. A positive weight can be absorbed into an invertible linear reparameterization. The result does not cover general covariance parameters.

Comment

The unresolved question is a general criterion for equality with the restricted single-copy measurement cost. A finite optimization of the Holevo bound already exists. Quadratic optimizing observables can require non-Gaussian detection, and collective asymptotic attainability is a different property.

References

[CGA26]
S. Chang, M. G. Genoni, and F. Albarelli, “Efficiently Evaluating Holevo, RLD and SLD Cramér-Rao Bounds for Multiparameter Quantum Estimation with Gaussian States,” Communications Physics 9, 126 (2026).DOIarXiv
[BLA18]
M. Bradshaw, P. K. Lam, and S. M. Assad, “Ultimate Precision of Joint Quadrature Parameter Estimation with a Gaussian Probe,” Physical Review A 97, 012106 (2018).DOIarXiv

Page edit log

  • Record created
  • Last edited
  • Revisions1

View the full history on GitHub

Your contribution is welcome!

Found progress, a correction, or a resolution? Edit this record on GitHub and open a pull request, or report an update with the primary sources. The proposal page explains the available submission route; see the contribution guide for details.

Cite this page

“Gaussian-measurement equality with the Holevo bound,” Quantum Information and Quantum Computation Open Problem Zoo (QIQCOP Zoo), ID op_8f1853475db7ea27, accessed 2026-09-16.

Use the Cite button above for BibTeX and the permanent link.

Cite this problem

Please also cite the primary sources listed under References. Cite this page for the statement, status, and stable identifier.

BibTeX

@incollection{qiqcop_op_8f1853475db7ea27,
  title = {Gaussian-measurement equality with the Holevo bound},
  booktitle = {Quantum Information and Quantum Computation Open Problem Zoo (QIQCOP Zoo)},
  year = {2026},
  howpublished = {\url{https://qiqc-op.com/problem/op_8f1853475db7ea27/}},
  note = {Stable ID op_8f1853475db7ea27; status: Unsolved; accessed 2026-09-16}
}

Plain text

“Gaussian-measurement equality with the Holevo bound,” Quantum Information and Quantum Computation Open Problem Zoo (QIQCOP Zoo), https://qiqc-op.com/problem/op_8f1853475db7ea27/, ID op_8f1853475db7ea27, accessed 2026-09-16.

Share this problem

Permanent link

Identifiers

op_8f1853475db7ea27
01M26K8Q8DCN2WF4YHQ0RJX373