Pith. sign in

REVIEW 3 major objections 5 minor 19 references

Conditions for Quantum Advantage in AC Power Flow

T0 review · 3 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper establishes that any gate-based quantum AC power flow solver with full classical readout is slower than Newton-Raphson at every practical accuracy, because its runtime depends on one over the error tolerance instead of the…

desk verdict Useful negative result for quantum ACPF under an explicit full-tomography pipeline, but the abstract's 'any gate-based algorithm' claim overreaches; worth refereeing after tightening. read the letter →

arxiv 2608.06711 v1 pith:PINTSMA4 submitted 2026-08-07 quant-ph cs.SYeess.SY

Parikshit Pareek , Abhijith Jayakumar , Carleton Coffrin , Sidhant Misra This is my paper · ORCID
classification quant-phcs.SYeess.SY
keywords ACpowerflowquantumadvantagelinearsystemsNewton-Raphsonloadruntimecomplexitytomographyreadouterrortolerance
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks whether a quantum computer can solve alternating current power flow (ACPF) faster than the classical Newton-Raphson load flow (NRLF). It argues that the standard end-to-end recipe—prepare the load vector as a quantum state, run a quantum linear-system solver, then read out the dense voltage-and-angle solution classically—cannot beat NRLF at any practically useful accuracy. The decisive term is readout: recovering the $N$-bus solution to error $\varepsilon$ costs $\Theta(N/\varepsilon)$, and because a measurement destroys the quantum state, this cost is paid after every solve, giving a total lower bound $\Omega(N\kappa/\varepsilon)$. NRLF, by contrast, costs $O(N\kappa \log(\kappa/\varepsilon))$. Since $1/\varepsilon$ outruns $\log(\kappa/\varepsilon)$, the quantum–classical ratio only falls below one near $\varepsilon \approx 0.05$, so the paper concludes that quantum advantage for ACPF, if any, must come from tasks other than direct dense solution, or from regimes where classical overheads are much larger.

What carries the argument

The machinery is the three-stage end-to-end decomposition of a quantum ACPF solver plus the error-norm comparison that puts both solvers on the same footing. State preparation $T_p$ encodes the injection vector by amplitude embedding ($\Theta(N)$ generally, $\Theta(\log N)$ with QRAM); state propagation $T_s$ is the quantum linear-solve step, bounded below by $\Omega(\kappa)$ from HHL-style query complexity; readout $T_r$ is quantum tomography with state-preparation unitaries, costing $\Theta(N/\varepsilon)$ to recover a dense classical voltage vector in $\ell^2$ norm. Because a single measurement yields one sample and collapses the state, the full preparation-and-solve loop runs $T_r$ times, and the product $T_r(T_p+T_s)$ produces the $\Omega(N\kappa/\varepsilon)$ lower bound. On the classical side, the key mechanism is CG on the normal equations $A^T A x = A^T b$, which squares the condition number, and the conversion from energy-norm error $\varepsilon_c$ to $\ell^2$ error $\varepsilon$ yields $O(N\kappa \log(\kappa/\varepsilon))$.

What would settle it

Run an end-to-end quantum ACPF implementation on a fixed test case at two error targets, say $\varepsilon = 10^{-3}$ and $\varepsilon = 10^{-6}$; the claimed readout term predicts about a $10^3$-fold runtime increase, whereas a logarithmic dependence predicts only about a factor of two. A measured increase close to two, or any solver achieving total runtime $O(N\kappa \log(\kappa/\varepsilon))$ at $\varepsilon = 10^{-6}$, would falsify the paper's lower bound.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is a baseline lower bound for any gate-based, end-to-end quantum ACPF algorithm. Splitting the run into state preparation $T_p = \Theta(\log N)$ (with QRAM), state propagation $T_s = \Omega(\kappa)$ (optimistically $\Theta(\kappa)$), and tomography readout $T_r = \Theta(N/\varepsilon)$, and noting that measurement destroys the state so all three stages must be repeated $T_r$ times, the end-to-end runtime is $\Theta(T_r(T_p+T_s)) = \Omega(N\kappa/\varepsilon)$. The comparison case is NRLF solved with conjugate gradients on the normal equations, whose per-iteration cost is $O(N\kappa \log(\kappa/\varepsilon))$. The ratio $Q/C \approx (1/\varepsilon)/\log(\kappa/\varepsilon)$ therefore does not depend on $N$; at $\kappa = 10^8$ it is about $3.1 \times 10^4$ when $\varepsilon = 10^{-6}$ and falls below one only near $\varepsilon \approx 0.05$. The authors conclude that potential quantum benefit should be sought in tasks other than direct dense ACPF solution—enumerating multiple solutions, bifurcation detection, or convergence robustness—and only if those variants beat classical counterparts such as Holomorphic-NRLF.

Load-bearing premise

The load-bearing premise is that any useful quantum ACPF solver must classically read out the full dense solution vector, with readout cost proportional to the number of buses divided by the error tolerance; if a solver only reports a few operator-relevant numbers, or hands the quantum state to another quantum routine, the claimed lower bound no longer applies.

Editorial extensions

If this is right

  • At error tolerance $10^{-6}$ and condition number $10^8$, the quantum baseline is about $3.1 \times 10^4$ times the classical cost; at $10^{-3}$ the ratio is about 39.
  • The quantum–classical ratio is independent of system size $N$, so simply moving to larger power grids does not create a quantum advantage under this model.
  • The quantum solver's advantage window opens only near $\varepsilon \approx 0.05$ for $\kappa = 10^8$, an error too large for operational power-flow studies.
  • The logarithmic error dependence of HHL inside the linear solve is erased by the $1/\varepsilon$ tomography cost, so a quantum ACPF speedup cannot rest on the linear-solve speedup alone.
  • The paper's remaining positive directions are nonlinear tasks—enumerating multiple power-flow solutions, detecting bifurcations and ill-conditioning, and improving initialization robustness—provided those quantum routines beat classical alternatives such as Holomorphic-NRLF on the same task.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • This is an inference: the same readout barrier applies to any quantum algorithm whose final deliverable is a dense classical vector, not just power flow; quantum solvers for other nonlinear systems would face the same one-over-error wall.
  • This is an inference: a quantum ACPF routine that reports only operator-relevant scalars—line flows, total losses, constraint violations—via amplitude estimation could escape the bound; the paper's lower bound explicitly assumes full tomography of the solution vector.
  • This is an inference: because the quantum/classical ratio is independent of system size, the only paths to advantage are algorithmic (changing what is read out or which nonlinear task is solved), not architectural (waiting for bigger grids).
  • This is an inference: replacing the full-tomography assumption with a few-output query model and recomputing the end-to-end complexity is a direct test of how much of the conclusion depends on readout; one would predict a bound that scales with the number of reported quantities rather than with $N$.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper analyzes the runtime complexity of quantum algorithms for AC power flow (ACPF). It first establishes an NRLF benchmark complexity of O(N κ log(κ/ε)), with N buses, condition number κ, and error tolerance ε. It then attempts to derive an 'optimistic lower bound' of Ω(Nκ/ε) for any gate-based quantum ACPF solver, determined by state preparation, a quantum linear solve, and dense tomography readout. The resulting Q/C complexity ratio is approximately 1/(ε log(κ/ε)), leading the authors to conclude that no practical quantum advantage exists except at very coarse accuracies or with large classical overheads.

Significance. If the lower bound were valid, the paper would provide an important negative result for quantum power flow research. The paper is valuable in systematically assembling the NRLF complexity benchmark, explicitly converting between energy-norm and ℓ2-norm errors, and supplying empirical condition-number data from PGLib. It also clearly identifies the 1/ε readout bottleneck as the key obstacle. However, the central 'any gate-based QC algorithm' claim is not supported by the derivation, which analyzes one specific pipeline rather than proving a universal lower bound. The contribution is best read as a careful cost estimate for a canonical quantum ACPF pipeline; in its current form, the abstract overstates the scope of the conclusion.

major comments (3)
  1. [Section III (readout complexity) and Fig. 2] The abstract's claim of Ω(Nκ/ε) for 'any Gate-based QC algorithm' is not established. The derivation assigns T_r = Θ(N/ε) by requiring full dense tomography of the 2N-dimensional solution vector via the algorithm of [18]. No argument rules out a quantum ACPF algorithm that outputs only operator-relevant quantities—total active power loss, line-flow violations, an objective value—using amplitude estimation, whose readout cost would be O(1/ε) independent of N. Such an algorithm would have end-to-end cost roughly O((κ+log N)/ε), removing the N from the advantage ratio. The 'any' quantification therefore applies to the specific pipeline in Fig. 2, not to all possible quantum solvers.
  2. [Section III (state propagation and readout)] The terms 'lower bound' and 'tight bound' are used in a way that is not justified. Reference [18] provides a tomography algorithm with cost Θ(N/ε); this is an upper bound on one readout strategy, and no matching lower bound is proved. Similarly, the Ω(κ) query lower bound for quantum linear solvers is for a black-box matrix access model and does not imply that every ACPF algorithm must incur this cost for the specific Jacobian structure. The assembled expression is a cost estimate for one architecture, not a fundamental complexity bound.
  3. [Fig. 2 and 'Quantum Advantage Regime'] The comparison assumes H=1 for the quantum ACPF solver while treating the NRLF iteration count H as a constant. Since ACPF is a nonlinear problem, a single linear solve does not generally produce the solution; the assumption is explicitly optimistic, but it is an assumption rather than a lower bound. The paper does not discuss how the complexity ratio would change if the quantum algorithm also requires multiple outer iterations with possibly growing condition numbers. This affects the central conclusion about the absence of practical advantage.
minor comments (5)
  1. [Section II, equation (3)] The derivation of O(N κ log(κ/ε)) from (2) relies on the energy-norm to ℓ2-norm error conversion. This step is only sketched in the text; elaborating the chain ε_c ≥ ε/κ would improve clarity.
  2. [Appendix A] The numbered equality chain (5)–(8) is misaligned: the items with equality signs appear as separate equations rather than as a connected derivation. Please rewrite as a single aligned chain or with explicit inequality labels.
  3. [Section II and Fig. 1] The claim that the iteration count H is independent of system size is supported only by the left panel for converging instances. The right panel shows non-converging instances, which are not used for the benchmark; the text should state this distinction explicitly.
  4. [Footnote 1] The footnote contains the placeholder 'Link' for big-O notation references; this should be replaced with a proper citation or removed.
  5. [Quantum Advantage Regime] The Q/C ratio is computed as 1/(ε log(κ/ε)) after canceling the common factor N. The paper should state explicitly that this assumes identical constant factors for both algorithms, because the qualitative conclusion depends on that assumption.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity; the central Ω(Nκ/ε) bound is assembled from external complexity results, with only minor non-load-bearing self-citations, though the full-tomography universality assumption is a correctness gap.

full rationale

The paper's derivation chain is self-contained against external benchmarks. The classical NRLF bound O(Nκ log(κ/ε)) follows from standard CG complexity applied to the normal equations, with κ taken from PGLib condition-number scans; no step defines the conclusion in terms of its own inputs. The quantum "lower bound" Ω(Nκ/ε) is assembled from three cited components: amplitude-encoding state preparation (Θ(log N) with QRAM, from [17]), the quantum linear-solver condition-number scaling (Ω(κ), attributed to [9]), and tomography via state-preparation unitaries (Θ(N/ε), from [18]). The end-to-end product T_r(T_p+T_s) is a complexity estimate of the specific pipeline in Fig. 2, not a mathematically forced lower bound for every conceivable output model; an algorithm that returns only a few expectation values could avoid the N/ε readout. This is an unproven universality assumption, i.e., a correctness risk, not circularity, because the paper does not fit parameters to data and then "predict" that same data, nor does it cite its own prior work to forbid alternatives. Self-citations to [10] are used for context (pipeline figure, sparsity and condition-number observations) but are not load-bearing for the central asymptotic comparison, which rests on external references [9], [17], [18]. Accordingly, the honest finding is no significant circularity; score 2 reflects only the presence of minor non-load-bearing self-citations.

Assumptions & free parameters 0 free parameters · 8 assumptions · 0 invented entities

The central comparison uses no fitted parameters. It rests on an empirical classical-benchmark assumption (H and sparsity constant), an idealized quantum pipeline (QRAM, single iteration, full tomography), and standard complexity bounds for CG, QLSA, and quantum state tomography. The universal 'any gate-based algorithm' wording is broader than the pipeline actually analyzed.

assumptions (8)
  • domain assumption NRLF iteration count H is independent of system size N.
    Section II drops H from O(H N s κ log(...)) to O(Nκ log(...)) using Fig. 1 and [1]; the empirical constancy of H is load-bearing for the classical benchmark.
  • domain assumption Jacobian sparsity s is bounded by a constant across PGLib systems.
    Section II cites [10] that Y-Bus sparsity stays below 30, removing s from the complexity; this makes the classical benchmark tighter.
  • ad hoc to paper The quantum pipeline consists of amplitude-encoded state preparation, a quantum linear solve, and dense tomography readout.
    Section III assumes this three-component pipeline for any gate-based algorithm; not proven to be universal, so the abstract's 'any Gate-based QC algorithm' claim is stronger than the derivation.
  • ad hoc to paper QRAM is available so state preparation costs Θ(log N).
    Section III uses QRAM as an optimistic assumption; without it, T_p=Θ(N) and the quantum bound is even worse.
  • ad hoc to paper A single outer iteration suffices for the quantum ACPF solve (H=1).
    Fig. 2 sets H=1 as an optimistic lower bound for the quantum side; this is favorable to quantum but not justified by any algorithm.
  • domain assumption Quantum linear solvers have query complexity lower bound Ω(κ).
    Section III cites [9] and takes T_s=Θ(κ); standard for black-box QLSA, but not an unconditional time lower bound for all encodings.
  • domain assumption Tomography with state-preparation unitaries costs Θ(N/ε).
    Section III cites [18] for T_r=Θ(N/ε); assumes access to the preparing unitary and ignores potential extra query costs.
  • standard math CG on normal equations gives O(N s κ log(1/ε_c)) for each NRLF linear solve.
    Section II equations (1)-(2); standard conjugate gradient complexity with κ(M)=κ(A)^2.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Conditions for Quantum Advantage in AC Power Flow." pith.science (2026). https://pith.science/paper/PINTSMA4

@misc{pith2026260806711,
  author       = {Pith},
  title        = {Pith review of: Conditions for Quantum Advantage in AC Power Flow},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PINTSMA4}},
  note         = {Machine review of arXiv:2608.06711}
}
abstract

This paper aims to contextualize the requirements for Quantum Computing (QC) algorithms to achieve a quantum advantage in solving the alternating current power flow (ACPF) problem, with a focus on runtime complexity. First, we establish a benchmark for a QC iterative solver to demonstrate an advantage over the classical Newton-Raphson Load Flow (NRLF) algorithm. Next, we derive a baseline expression for the end-to-end runtime complexity of any Gate-based QC algorithm as $\Omega(N \kappa/\varepsilon),$ reflecting dependence on system size $N$, condition number $\kappa$, and error tolerance $\varepsilon$. Finally, we highlight key areas where QC algorithms may offer potential benefits over NRLF in addressing the standard ACPF problem.

Figures

Figures reproduced from arXiv: 2608.06711 by the authors.

Figure 1
Figure 1. Condition number (κ) variations for the PGLib [13] power transmis￾sion network datasets reveal potential numerical challenges in solving ACPF. Left: Variation in the Jacobian condition number in converging instances for medium-scale systems. It is clear that the condition number increases with each iteration before converging to the final value. Right: Variation in the Jacobian condition number across 20 NRLF iterat… view at source ↗
Figure 2
Figure 2. Top: NRLF complexity bound considering independence of sparsity and iterations H with respect to system size N and error bounds. Bottom: Lower bound of quantum complexity considering QRAM availability, state propagation, readout and optimistic consideration of single iteration solve. tion vector state3 . The runtime complexity for this operation is Tp = Θ(N), considering complexity for preparing a general quantum st… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

19 extracted references · 16 canonical work pages

  1. [10]

    Demystifying quan- tum power flow: Unveiling the limits of practical quantum advantage,

    P. Pareek, A. Jayakumar, C. Coffrin, and S. Misra, “Demystifying quan- tum power flow: Unveiling the limits of practical quantum advantage,” arXiv preprint arXiv:2402.08617, 2024

    work page arXiv 2024

  2. [18]

    Quan- tum tomography using state-preparation unitaries,

    J. van Apeldoorn, A. Cornelissen, A. Gily ´en, and G. Nannicini, “Quan- tum tomography using state-preparation unitaries,” inProceedings of the 2023 Annual ACM-SIAM Symposium on Discrete Algorithms (SODA). SIAM, 2023, pp. 1265–1318

    work page 2023

  3. [1]

    J. J. Grainger and W. D. Stevenson,Power system analysis. McGraw- Hill, 1994

    work page 1994

  4. [2]

    A critical review on probabilistic load flow studies in uncertainty constrained power systems with photovoltaic generation and a new approach,

    B. R. Prusty and D. Jena, “A critical review on probabilistic load flow studies in uncertainty constrained power systems with photovoltaic generation and a new approach,”Renewable and Sustainable Energy Reviews, vol. 69, pp. 1286–1302, 2017

    work page 2017

  5. [3]

    Existing approaches and trends in uncertainty modelling and probabilistic stability analysis of power systems with renewable generation,

    K. N. Hasan, R. Preece, and J. V . Milanovi ´c, “Existing approaches and trends in uncertainty modelling and probabilistic stability analysis of power systems with renewable generation,”Renewable and Sustainable Energy Reviews, vol. 101, pp. 168–180, 2019

    work page 2019

  6. [4]

    Quantum power flow,

    F. Feng, Y . Zhou, and P. Zhang, “Quantum power flow,”IEEE Transac- tions on Power Systems, vol. 36, no. 4, pp. 3810–3812, 2021

    work page 2021

  7. [5]

    Quantum computing for enhancing grid security,

    R. Eskandarpour, P. Gokhale, A. Khodaeiet al., “Quantum computing for enhancing grid security,”IEEE Transactions on Power Systems, vol. 35, no. 5, pp. 4135–4137, 2020

    work page 2020

  8. [6]

    Quantum computing for power flow algorithms: Testing on real quantum computers,

    B. Sævarsson, S. Chatzivasileiadiset al., “Quantum computing for power flow algorithms: Testing on real quantum computers,” in11th Bulk Power Systems Dynamics and Control Symposium, 2022

    work page 2022

Show all 19 references
  1. [7]

    Quantum power flows: From theory to practice,

    J. Liu, H. Zheng, M. Hanada, K. Setia, and D. Wu, “Quantum power flows: From theory to practice,”arXiv preprint arXiv:2211.05728, 2022

    2022 arXiv

  2. [8]

    Quantum computation in power systems: An overview of recent advances,

    S. Golestan, M. Habibi, S. M. Mousaviet al., “Quantum computation in power systems: An overview of recent advances,”Energy Reports, vol. 9, pp. 584–596, 2023

    2023

  3. [9]

    Quantum algorithm for linear systems of equations,

    A. W. Harrow, A. Hassidim, and S. Lloyd, “Quantum algorithm for linear systems of equations,”Physical review letters, vol. 103, no. 15, p. 150502, 2009

    2009

  4. [11]

    Milano,Power system modelling and scripting

    F. Milano,Power system modelling and scripting. Springer Science & Business Media, 2010

    2010

  5. [12]

    An introduction to the conjugate gradient method without the agonizing pain,

    J. R. Shewchuk, “An introduction to the conjugate gradient method without the agonizing pain,” 1994

    1994

  6. [13]

    The power grid library for benchmarking ac optimal power flow algorithms,

    S. Babaeinejadsarookolaeeet al., “The power grid library for benchmarking ac optimal power flow algorithms,”arXiv preprint arXiv:1908.02788, 2019

    1908 arXiv

  7. [14]

    Powerworld simulator,

    “Powerworld simulator,” https://www.powerworld.com/knowledge-base/ what-power-flow-algorithm-does-powerworld-use

  8. [15]

    Comparison of the holomorphic embedding load flow method with established power flow algorithms and a new hybrid approach,

    P. S. Sauter, C. A. Braun, M. Kluwe, and S. Hohmann, “Comparison of the holomorphic embedding load flow method with established power flow algorithms and a new hybrid approach,” in2017 Ninth Annual IEEE Green Technologies Conference (GreenTech), 2017, pp. 203–210

    2017

  9. [16]

    Schuld and F

    M. Schuld and F. Petruccione,Supervised learning with quantum com- puters. Springer, 2018, vol. 17

    2018

  10. [17]

    Quantum algorithms: A survey of applications and end-to-end complexities,

    A. M. Dalzellet al., “Quantum algorithms: A survey of applications and end-to-end complexities,”arXiv preprint arXiv:2310.03011, 2023

    2023 arXiv

  11. [19]

    Numerical polynomial homotopy continuation method to locate all the power flow solutions,

    D. Mehtaet al., “Numerical polynomial homotopy continuation method to locate all the power flow solutions,”IET Generation, Transmission & Distribution, vol. 10, no. 12, pp. 2972–2980, 2016. APPENDIX Notice that quantum and classical linear system solvers solve different system...

    2016

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.