Electric Power Demand Portfolio Optimization by Fermionic QAOA with Self-Consistent Local Field Modulation

Keisuke Fujii; Keita Sasada; Takuya Yoshioka; Yuichiro Nakano

REVIEW 4 major objections 5 minor 25 references

Electric Power Demand Portfolio Optimization by Fermionic QAOA with Self-Consistent Local Field Modulation

T0 review · 4 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read A fermionic QAOA variant embeds the power-procurement constraint in its driver Hamiltonian and, in noiseless simulation, achieves lower expected cost than two standard QAOA baselines on all eight demand periods tested.

desk verdict A real new QAOA driver mechanism, but the reported advantage is measured against a penalized proxy that the paper itself concedes can violate the actual feasibility constraint. read the letter →

arxiv 2505.02282 v1 pith:A6MNHOFO submitted 2025-05-04 quant-ph

Takuya Yoshioka , Keita Sasada , Yuichiro Nakano , Keisuke Fujii This is my paper · ORCID

classification quant-ph

keywords electricpowerdemandportfoliooptimizationnegawatttradingFermionicQAOAself-consistentlocalfieldmodulationHartree-FockdriverHamiltonianresponseconstrainedquantum

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

The paper proposes FQAOA-SCLFM, a version of the Fermionic Quantum Approximate Optimization Algorithm that builds the constraint on the target amount of procured power directly into the driver Hamiltonian, rather than leaving it only to the cost function. It does this by replacing the quadratic penalty on the total-negawatt deviation with a linear local field whose strength is fixed self-consistently by a Hartree-Fock calculation, so the driver's ground state is a single Slater determinant and can be prepared efficiently. Applied to an electricity demand portfolio problem—choosing which demand-response participants to call so that total negawatt matches the desired procurement with low variance—the method yields lower expected cost than XY-QAOA and the previous FQAOA in every tested demand period at QAOA levels $p=0$, $1$, and $10$ in noiseless simulation. If this holds outside the simulated instances, aggregators could procure negawatt portfolios that meet the target more reliably with the same or fewer quantum operations.

What carries the argument

Self-consistent local field modulation (SCLFM): the construction of a free-fermion driver Hamiltonian $$\hat{H}^{\mathrm{HF}}_{d,T} = -t_{\mathrm{hop}}\sum_l (\hat{c}^{\dagger}_{l+1}\hat{c}_l + \mathrm{h.c.}) + \frac{2}{N_T}\sum_t ($P^{{\mathrm{HF}}$}_{t,\mathrm{tot}} - P'_{t,\mathrm{proc}})\sum_l \mathbb{E}[p_{t,l}]\hat{n}_l - \frac{1}{N_T}\sum_t\left[($P^{{\mathrm{HF}}$}_{t,\mathrm{tot}})^2 - (P'_{t,\mathrm{proc}})^2\right],$$ with $P^{\mathrm{HF}}_{t,\mathrm{tot}}=\sum_l \mathbb{E}[p_{t,l}]\langle \phi_{\mathrm{HF}}|\hat{n}_l|\phi_{\mathrm{HF}}\rangle$ computed self-consistently from the driver's own ground state. This replaces the quadratic soft constraint by a linear local field while keeping the Hamiltonian quadratic in fermionic operators, so the ground state remains a single Slater determinant satisfying the particle-number constraint. The same Hartree-Fock iteration yields the initial state, and the mixing unitary factorizes into number, boundary, odd, and even parts whose implementation adds only $pL$ single-qubit Pauli-$Z$ gates to the previous FQAOA circuit.

What would settle it

Run the same noiseless comparison at a procurement target deliberately far from the range of achievable totals (for example, set $P'_{t,\mathrm{proc}}$ to the maximum achievable total with $M_T=5$, instead of $1.5\,\mathrm{kWh}$): if FQAOA-SCLFM's expected cost at $p=1$ no longer beats the previous FQAOA, the SCLFM advantage is conditional on the Hartree-Fock approximation's accuracy rather than intrinsic to the driver construction.

Watch

Extended reading notes

Core claim

The central discovery, on the paper's own terms, is that the self-consistent local field modulation removes a known defect of the previous FQAOA driver: the soft constraint $(1/N_T)\sum_t (\hat{P}_t - P'_{t,\mathrm{proc}})^2$ is quadratic in the number operator, so its ground state is not a single Slater determinant and the driver violates the design condition that the initial state be efficiently preparable. The proposed driver keeps only the linear part of this penalty, evaluated at a Hartree-Fock fixed point, and solves the resulting free-fermion Hamiltonian self-consistently. In noiseless simulations on a 20-participant, five-request problem fixed to $P'_{t,\mathrm{proc}}=1.5\,\mathrm{kWh}$ across eight three-hour periods, the resulting algorithm reaches expected cost (Eq. 26) below both XY-QAOA and previous FQAOA at every QAOA level tested; at $p=10$ the total negawatt $P_{t,\mathrm{tot}} \pm \sigma_{t,\mathrm{tot}}$ roughly satisfies the original balance inequality, and at $p=0$ it already sits near the target in the difficult evening period. The additional circuit cost is only $pL$ single-qubit $Z$ rotations.

Load-bearing premise

The load-bearing premise is that the Hartree-Fock estimate of the total procured power is close enough to the true optimum that the quadratic procurement penalty can be replaced by a linear local field, and that the penalized cost function is an acceptable stand-in for the original balance inequality even though it can violate it.

Editorial extensions

If this is right

FQAOA-SCLFM reports lower expected cost than both XY-QAOA and the previous FQAOA in all eight demand periods at QAOA levels $p=0$, $1$, and $10$, so on this problem it is the better performing of the three algorithms at the depths tested.
At $p=10$, the optimized state yields total negawatt $P_{t,\mathrm{tot}}\pm\sigma_{t,\mathrm{tot}}$ that roughly satisfies the balance condition, meaning procurement risk is reduced enough for near-target delivery.
Because the SCLFM driver is derived from the soft constraint, the same ansatz construction applies to other constrained combinatorial optimizations whose cost has one hard cardinality constraint and one quadratic soft penalty.
The added resource cost is only $pL$ single-qubit Pauli-$Z$ rotations, so the improvement does not require deeper circuits than the previous FQAOA.
The $T=18$ and $T=21$ instances, which have large positive and negative negawatt covariances, show the largest relative improvement, suggesting the benefit is most pronounced on strongly fluctuating demand periods.

Reading between the lines

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

One consequence the authors leave implicit is that most of the advantage is likely coming from the initial state: already at $p=0$ the Hartree-Fock state puts the total negawatt near the target, so a cleaner test would be to keep the SCLFM initial state but swap in the standard FQAOA driver during mixing, isolating the contribution of the driver itself.
The results cover one operating point ($P'_{t,\mathrm{proc}}=1.5\,\mathrm{kWh}$, $M_T=5$, $L=20$) and one residential dataset; a testable extension is to vary the procurement target, particularly to values far from the Hartree-Fock solution, where the linearization of the quadratic penalty should lose accuracy.
On portfolios with many near-degenerate low-cost solutions, the Hartree-Fock fixed point may be a poor proxy for the true optimum; an adaptive scheme that re-estimates the local field from the QAOA output distribution rather than only from the Hartree-Fock state is a natural next step, but it is not what this paper proposes.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

The SCLFM driver is a genuine new mechanism: replacing the quadratic soft constraint in the driver with a self-consistent linear field is not a one-line tweak, and the HF fixed-point loop plus the extra U_I rotation make it a distinct contribution to the FQAOA literature. The derivation of Eq. (15) from Eq. (14) is algebraically fine, and the numerics in Table I are internally consistent: on the penalized cost E_T,x, FQAOA-SCLFM beats both XY-QAOA and the earlier FQAOA on all eight time windows at p=0,1,10.

That said, the paper's own text concedes the central weakness: Eq. (4) "guarantees efficiency and feasibility ... while violating the inequality constraint in Eq. (3)." All reported costs, including the E_T,min baseline in Eq. (26), are measured against this penalized proxy, not against the original variance-minimization-with-inequality problem. So we do not actually know whether the lower ΔE_T values correspond to portfolios that are feasible or closer to the true constrained optimum. The stress-test note is on target here, and the paper does not offer an infeasibility rate or any argument that the unit penalty weight makes the proxy a good Lagrangian relaxation.

Other soft spots, in proportion: the demonstration is tiny (20 variables, noiseless simulation only), there are no error bars or multiple optimization runs, and no code or data is shipped (though the input data is from a cited public database). There is also no serious classical baseline—random sampling under the constraint appears only as an inset in Fig. 4, and a simple local search or greedy would likely do well on these instances. For an algorithmic paper these gaps are survivable, but they undercut any practical claim about stable negawatt procurement.

What is worth keeping: the idea of using the soft-constraint structure to build a problem-adapted driver, and solving it self-consistently at the mean-field level, is clever and could transfer to other soft-constrained QUBO-like problems. The p=0 advantage is exactly what you'd expect from a mean-field initialization that already respects the bias toward P'_t,proc, and the paper's explicit acknowledgment of the Eq. (3) violation is to its credit.

My take: this deserves a serious referee, not a desk reject—the mechanism is novel and the math is sound. But the referee should push for feasibility tracking, a classical baseline, and a statement about what happens to the advantage when Eq. (3) is actually enforced. As it stands, I'd read it as a contribution to QAOA driver design, not as evidence about power portfolio optimization.

Referee Report

4 major / 5 minor

Summary. The manuscript proposes FQAOA-SCLFM, a fermionic QAOA variant whose driver Hamiltonian is augmented by a self-consistent local field obtained from a Hartree-Fock approximation of the soft target-power constraint. The method is applied to an electricity demand portfolio problem in which a binary vector x selects M DR participants and the objective is to procure a target negawatt with small variance. The paper introduces a penalized quadratic cost function, maps it to a fermionic Hamiltonian, linearizes the penalty term in the driver through Eq. (15), and reports noiseless simulations on eight time windows using a 20-participant dataset. Table I reports that FQAOA-SCLFM achieves lower expected penalized costs than XY-QAOA and the previous FQAOA at QAOA levels p=0, 1, and 10 across all eight windows.

Significance. The algebraic derivation of Eq. (15) as a linearized, self-consistent version of the driver in Eq. (14) is straightforward and appears correct, and the paper provides a complete grid of simulation results over all time windows and three QAOA levels. If the reported advantage survives evaluation against the original inequality-constrained problem, the SCLFM construction is a useful and inexpensive way to encode a soft constraint into the QAOA driver and initial state, adding only pL extra single-qubit rotations. The main limitations are that the benchmark is a surrogate objective with no feasibility statistics, the simulations lack repeated-run statistics, and the instance set is narrow; these limitations currently make the significance conditional.

major comments (4)

[Section II, Eq. (4) and Table I] The problem stated in Eqs. (1)-(3) is variance minimization under the inequality constraint P_{t,proc} <= sum_l E[p_{t,l}] x_l <= P_{t,proc} + delta, but the quantity minimized and reported is the unit-weight penalized proxy E_{t,x} of Eq. (4). Section II itself concedes that this proxy 'violat[es] the inequality constraint in Eq. (3)', yet Table I reports only Delta E_T relative to E_{T,min}, the minimum of the proxy. A lower Delta E_T therefore does not by itself establish better performance on the actual procurement objective. Please report the fraction of sampled portfolios that violate Eq. (3), the distribution of sum_l E[p_{t,l}] x_l for each algorithm, and the expectation of the original variance under the constraint (or a bound quantifying the suboptimality introduced by the unit-weight penalty). Without such data, the claim that FQAOA-SCLFM procures negawatt 'more stably' is not supported.
[Section IV.B.2, Table I] Table I reports a single value per algorithm, instance, and QAOA level, with no description of the number of BFGS restarts, random seeds, or confidence intervals. Several of the claimed advantages are small enough to be optimizer noise, e.g. T=6, p=1 (XY-QAOA 0.060 vs FQAOA 0.056), T=9, p=10 (FQAOA 0.037 vs FQAOA-SCLFM 0.036), and T=0, p=1 (XY-QAOA 0.073 vs FQAOA 0.066). Since the central claim is uniform outperformance over all instances, the paper should provide repeated optimization runs, report best/median statistics, and give error bars or a statistical test.
[Section IV.B.2, text after Table I] The sentence 'For any given T and finite p, FQAOA outperforms XY-QAOA, and FQAOA-SCLFM further outperforms FQAOA' is contradicted by Table I: at p=0, T=3, FQAOA has Delta E_T/W_T = 0.123 versus 0.117 for XY-QAOA, and at p=0, T=12 the two values are equal at 0.177. Please correct this statement or restrict it to the actual central claim, namely that FQAOA-SCLFM outperforms both baselines in the table.
[Section III.B, Eq. (15)] The HF driver is obtained by replacing (P_t - P^HF_{t,tot})^2 with zero, but the quality of this approximation is not checked. Since the p=0 advantage of FQAOA-SCLFM appears to come from the HF initial state being close to the proxy optimum, the paper should quantify this closeness, for instance by comparing the HF state energy to the exact E_{T,min} on a few windows or by reporting overlap or energy-gap diagnostics. This is particularly important because E_{T,min} can be computed exactly for L=20, M=5 by brute force, yet the paper does not state how E_{T,min} was obtained.

minor comments (5)

[Appendix A, Eq. (A.2)] Equation (A.2) sets n^HF_{l,0} = 1/M_T, which does not describe a uniform distribution with M_T particles among L sites; the sum over l would be L/M_T, not M_T. This should presumably read n^HF_{l,0} = M_T/L, which is the occupation probability of the uniform superposition under the constraint.
[Section IV.A] The stated parameters are inconsistent: P'_{t,proc} = P_{t,proc} + delta/2 with P_{t,proc}=1 and P'_{t,proc}=1.5 implies delta = 1, not delta = 11 as written.
[Section III.D, Eqs. (19) and (22)] Equation (19) denotes the local-field unitary as U_n(beta), while Eq. (22) defines U_I(beta); please make the notation consistent.
[Section II, text after Eq. (4)] The phrase 'guarantees efficiency and feasibility ... while violating the inequality constraint in Eq. (3)' is self-contradictory; please rephrase to state clearly that Eq. (4) is an unconstrained penalized surrogate whose minimizer may be infeasible for Eq. (3).
[Section IV.B.2] The text contains minor language errors, e.g. 'selecton' should be 'selection' and 'is the difficult to solve' should be 'is difficult to solve'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the HF driver is a heuristic built from the problem's penalty structure, and all reported comparisons use the same cost function.

full rationale

The paper's central claim is an optimization performance comparison on a fixed penalized cost function (Eq. 4). The proposed FQAOA-SCLFM driver (Eq. 15) is obtained by linearizing the quadratic penalty term of the cost Hamiltonian (Eq. 10) around a self-consistently determined Hartree-Fock expectation value (Eq. 16). This is a standard mean-field approximation, not a circular reduction: the HF solution is computed from the problem data (E[p_t,l], covariances) and a uniform initial distribution, not from the known optimum of the target problem. The self-consistency condition is a fixed-point equation defining the driver, not the target result. The performance metric ΔE_T (Eq. 26) is the same penalized cost expectation for all methods, so the comparison is not a fitted quantity masquerading as a prediction. The only self-citations, Refs. [18] and [19], supply the baseline FQAOA ansatz and gate decompositions; these are implemented and compared on equal footing in noiseless simulation rather than invoked to force the SCLFM result. The paper's admission that the proxy cost (Eq. 4) can violate the original inequality constraint (Eq. 3) is a correctness-risk concern about the objective formulation, not a circularity in the derivation chain.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The algorithm introduces no new physical entities. The main uncharged assumptions are the HF linearization, the relaxed penalty cost, the representativeness of the historical consumption data, and the reliability of single-start BFGS. The two manual parameters thop and α are the only hand-tuned numbers beyond the problem instance definition.

free parameters (2)

thop (hopping rate in driver Hamiltonian, Eq. 15) = Not given; set so the hopping cost range equals the cost term range in Eq. (10)
Chosen by hand to match energy scales; affects the mixer strength and thus the QAOA dynamics.
α (mixing parameter for HF iterations, Eq. A.3) = Not stated in text; Fig. 5 shows α between 0.2 and 0.8, with α ≤ 0.6 stabilized
Numerical stabilization parameter, chosen ad hoc; convergence depends on it.

assumptions (4)

ad hoc to paper The quadratic soft-constraint term in Eq. (10) can be replaced by the self-consistent local-field linearization in Eq. (15).
The HF approximation (P_t - P^HF_t,tot)^2 ≈ 0 is introduced without a bound on its error; quality depends on the HF ground state being close to the constrained optimum.
domain assumption The alternative cost function Eq. (4) with penalty (sum E[p]x - P')^2 is an acceptable proxy for the original inequality constraint Eq. (3).
The paper states this cost 'violates the inequality constraint in Eq. (3)' (Sec. II), so feasibility is relaxed; the assumption is that optimizing the proxy still gives useful portfolios.
domain assumption The 20-residence electricity usage data from 2003 (Ref. [23]) represents negawatt supply distributions.
Model parameters σ and E[p] are estimated from this dataset as in Ref. [8]; the results are only as good as this proxy.
domain assumption BFGS parameter optimization (Sec. III-E) reliably finds near-global minima of the QAOA landscape.
No multi-start or global optimization is reported; the point estimates in Table I assume the found parameters are good enough.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Electric Power Demand Portfolio Optimization by Fermionic QAOA with Self-Consistent Local Field Modulation." pith.science (2026). https://pith.science/paper/A6MNHOFO

@misc{pith2026250502282,
  author       = {Pith},
  title        = {Pith review of: Electric Power Demand Portfolio Optimization by Fermionic QAOA with Self-Consistent Local Field Modulation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/A6MNHOFO}},
  note         = {Machine review of arXiv:2505.02282}
}

abstract

Quantum Approximation Optimization Algorithms (QAOA) have been actively developed, among which Fermionic QAOA (FQAOA) has been successfully applied to financial portfolio optimization problems. We improve FQAOA and apply it to the optimization of electricity demand portfolios aiming to procure a target amount of electricity with minimum risk. Our new algorithm, FQAOA-SCLFM, allows approximate integration of constraints on the target amount of power by utilizing self-consistent local field modulation (SCLFM) in a driver Hamiltonian. We demonstrate that this approach performs better than the currently widely used $XY$-QAOA and the previous FQAOA in all instances subjected to this study.

Figures

Figures reproduced from arXiv: 2505.02282 by the authors.

**Figure 1.** Time series data of expected negawatt E[pt,l] ± σt,l from DR participants l = 1 − 20, where σt,l is standard deviation of random variable pt,l. Here, according to Ref. [8], the electricity usage data for each of 20 residences in Ref. [23] is assumed to be the amount of negawatt that each participant can supply. 4 [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗

**Figure 2.** Mean covariance of negawatt σT ,l,l′ [(kWh)2] in Eq. (10) between DR participants l =1 − 20 in time periods T =3, 12, and 18. Here we follow the same procedure as in [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** Calculated results of total negawatt Pt,tot ±σt,tot of Eqs. (23) and (24) at each time t, where approximate optimal states |ψp(γ ∗, β ∗)⟩ are obtained from optimization for each time period T with (a) FQAOA and (b) FQAOA-SCLFM at p = 0, 1, and 10. Procurement of negawatt Pt,proc and P ′ t,proc are shown by solid and broken lines [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Probability distribution of cost D(E) at p = 1 integrated over the WT /10 range of Eq. (25) at T = 18. Cost expectation values ∆ET /WT of Eq. (26) shown in the inset, where we have shaded the region from the first quartile to the third quartile. These results are obtai…

**Figure 5.** Figure 5: Decrease of Hartree-Fock energy EHF T = ⟨ϕHF|HˆHF d,T |ϕHF⟩ using Eq. (15) from i = 0, and mean total negawatt P HF T ,tot = P t∈T P HF t,tot/NT using Eq. (16) in time period T = 18, where, α is mixing parameter in Eq. (A.3). The HF iteration at T = 18 for varying the …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 24 canonical work pages

[1]

Fundamentals and business model for resource aggregator of demand response in electricity markets

X. Lu, K. Li, H. Xu, F. Wang, Z. Zhou, and Y . Zhang, “Fundamentals and business model for resource aggregator of demand response in electricity markets” Energy, vol. 204, p. 117885, 2020

work page 2020
[2]

A summary of demand response in electricity markets

M.H. Albadi, E.F. El-Saadany, “A summary of demand response in electricity markets”, Electric Power Systems Research, vol. 78, pp. 1989–1996, 2008

work page 1989
[3]

Demand Side Management: Demand Response, Intelligent Energy Systems, and Smart Loads

P. Palensky and D. Dietrich, “Demand Side Management: Demand Response, Intelligent Energy Systems, and Smart Loads”, IEEE Trans- actions on Industrial Informatics, vol. 7, pp. 381-388, 2011

work page 2011
[4]

Demand response and smart grids—A survey

P. Siano, “Demand response and smart grids—A survey”, Renewable and Sustainable Energy Reviews, vol. 30, pp. 461-478, 2014

work page 2014
[5]

A comprehen- sive review on microgrid and virtual power plant concepts employed for distributed energy resources scheduling in power systems

S. M. Nosratabadi and R.-A. Hooshmand, E. Gholipour, “A comprehen- sive review on microgrid and virtual power plant concepts employed for distributed energy resources scheduling in power systems” Renewable and Sustainable Energy Reviews, vol. 67, pp. 341–363, 2017. a

work page 2017
[6]

Distributed Optimal Power Management via Negawatt Trading in Real-Time Electricity Market

Y . Okawa and T. Namerikawa, “Distributed Optimal Power Management via Negawatt Trading in Real-Time Electricity Market”, IEEE Trans. Smart Grid, vol. 8, pp. 3009-3019, 2017

work page 2017
[7]

Challenges and prospects for negawatt trading in light of recent technological developments

W. Tushar, T. K. Saha, C. Yuen, D. Smith, P. Ashworth, H. V . Poor and S. Basnet, “Challenges and prospects for negawatt trading in light of recent technological developments”, Nat. energy, vol. 5, pp. 834-841, 2020

work page 2020
[8]

A design method for electric power demand portfolio and its basic investigation -Application to design of aggregator portfolio-

T. Tsurumi and K. Tokoro, “A design method for electric power demand portfolio and its basic investigation -Application to design of aggregator portfolio-”, CRIEPI report, C17016 (2018) (in Japanese)

work page 2018

Show all 25 references

[9]

A design method for electric power demand portfolio and its basic investigation -Proposition of a method for a resource aggregator to procure negawatt effectively-

T. Tsurumi and K. Tokoro, “A design method for electric power demand portfolio and its basic investigation -Proposition of a method for a resource aggregator to procure negawatt effectively-”, CRIEPI report, C18005 (2019) (in Japanese)

2019
[10]

Portfolio optimiza- tion of electricity markets participation using forecasting error in risk formulation

R. Faia, T. Pinto, Z. Vale, and J. M. Corchado, “Portfolio optimiza- tion of electricity markets participation using forecasting error in risk formulation”, Electrical Power and Energy System vol. 129, p. 106739 2021

2021
[11]

Quantum computing for energy systems optimization: challenges and opportunities

A. Ajagekar and F. You, “Quantum computing for energy systems optimization: challenges and opportunities”, Energy, vol. 179, pp. 76-89, 2019

2019
[12]

Adapting Quantum Approximation Optimization Algorithm (QAOA) for Unit Commitment

S. Koretsky and P. Gokhale and J. M. Baker and J. Viszlai and H. Zheng and N. Gurung and R. Burg and E. Paaso and A. Khodaei and R. Eskandarpour and F. T. Chong, “Adapting Quantum Approximation Optimization Algorithm (QAOA) for Unit Commitment”, 2021 IEEE In- ternational Confe...

2021
[13]

Quantum distributed unit commitment: an application in microgrids

N. Nikmehr, P. Zhang, and M. A. Bragin, “Quantum distributed unit commitment: an application in microgrids”, IEEE Transactions on Power Systems, vol. 37, pp. 3592-3603, 2022

2022
[14]

A quantum approximate optimization algorithm

E. Farhi, J. Goldstone, and S. Gutmann, “A quantum approximate optimization algorithm”, arXiv:1411.4028 [quant-ph], 2014

2014 arXiv
[15]

From the quantum approximate optimization algorithm to a quantum alternating operator ansatz

S. Hadfield, Z. Wang, B. O’Gorman, E. G. Rieffel, D. Venturelli, and R. Biswas, “From the quantum approximate optimization algorithm to a quantum alternating operator ansatz”, Algorithms, vol. 12, p. 34, 2019

2019
[16]

Obstacles to Variational Quantum Optimization from Symmetry Protection

S. Bravyi, A. Kliesch, R. Koenig, and E. Tang, “Obstacles to Variational Quantum Optimization from Symmetry Protection”, Phys. Rev. Lett. vol. 125, p. 260505, 2020

2020
[17]

A review on Quantum Approximate Optimization Algorithm and its variants

K. Blekos, D. Brand, A. Ceschini, C.-H. Chou, R.-H. Li, K. Pandya, and A. Summer, “A review on Quantum Approximate Optimization Algorithm and its variants”, Physics Reports, vol. 1068, pp. 1-66, 2024

2024
[18]

Fermionic quantum approximate optimization algorithm

T. Yoshioka, K. Sasada, Y . Nakano, and K. Fujii, “Fermionic quantum approximate optimization algorithm”, Phys. Rev. Research vol. 5, p. 023071, 2023

2023
[19]

Experimental Demon- stration of Fermionic QAOA with One-Dimensional Cyclic Driver Hamiltonian

T. Yoshioka, K. Sasada Y . Nakano, and K. Fujii, “Experimental Demon- stration of Fermionic QAOA with One-Dimensional Cyclic Driver Hamiltonian”, 2023 IEEE International Conference on Quantum Com- puting and Engineering (QCE), vol. 1, 300-306, 2023

2023
[20]

XY mixers: analytical and numerical results for the quantum alternating operator ansatz

Z. Wang, N. C. Rubin, J. M. Dominy, and E. G. Rieffel, “ XY mixers: analytical and numerical results for the quantum alternating operator ansatz”, Phys. Rev. A, vol. 101, p. 012320, 2020

2020
[21]

Constrained quantum optimization for extractive summarization on a trapped-ion quantum computer

P. Niroula, R. Shaydulin, R. Yalovetzky, P. Minssen, D. Herman, S. Hu, and M. Pistoia, “Constrained quantum optimization for extractive summarization on a trapped-ion quantum computer”, Sci. Rep. vol. 12, p. 17171, 2022. arXiv:2206.06290v1 [quant-ph], 2022

2022 arXiv
[22]

Qulacs: a fast and versatile quantum circuit simulator for research purpose

Y . Suzuki, Y . Kawase, Y . Masumura, Y . Hiraga, M. Nakadai, J. Chen, K. M. Nakanishi, K. Mitarai, R. Imai, S. Tamiya, T. Yamamoto, T. Yan, T. Kawakubo, Y . O. Nakagawa, Y . Ibe, Y . Zhang, H. Yamashita, H. Yoshimura, A. Hayashi, and K. Fujii, “Qulacs: a fast and versatile qu...

2021
[23]

In this paper, we follow the study [8] and use the electricity usage of 20 residences in September 1-15, 2003 as the amount of negawatt

Database of energy consumption in residences is available at http://tkkankyo.eng.niigata-u.ac.jp/HP/HP/database/index.htm. In this paper, we follow the study [8] and use the electricity usage of 20 residences in September 1-15, 2003 as the amount of negawatt. Note that, electr...

2003
[24]

A Divide- and-Conquer Approach to Dicke State Preparation

S. Aktar, A. B ¨artschi, A.-H. A. Badawy, and S. Eidenbentz, “A Divide- and-Conquer Approach to Dicke State Preparation”, IEEE Trans. Quan- tum. Eng. vol. 3, 1-16, 2022

2022
[25]

Short-Depth Circuits for Dicke State Preparation

A. B ¨artschi and S. Eidenbenz, “Short-Depth Circuits for Dicke State Preparation”, 2022 IEEE International Conference on Quantum Com- puting and Engineering (QCE), pp. 87-96, 2022. 8

2022

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [1]

Fundamentals and business model for resource aggregator of demand response in electricity markets

X. Lu, K. Li, H. Xu, F. Wang, Z. Zhou, and Y . Zhang, “Fundamentals and business model for resource aggregator of demand response in electricity markets” Energy, vol. 204, p. 117885, 2020

work page 2020

[2] [2]

A summary of demand response in electricity markets

M.H. Albadi, E.F. El-Saadany, “A summary of demand response in electricity markets”, Electric Power Systems Research, vol. 78, pp. 1989–1996, 2008

work page 1989

[3] [3]

Demand Side Management: Demand Response, Intelligent Energy Systems, and Smart Loads

P. Palensky and D. Dietrich, “Demand Side Management: Demand Response, Intelligent Energy Systems, and Smart Loads”, IEEE Trans- actions on Industrial Informatics, vol. 7, pp. 381-388, 2011

work page 2011

[4] [4]

Demand response and smart grids—A survey

P. Siano, “Demand response and smart grids—A survey”, Renewable and Sustainable Energy Reviews, vol. 30, pp. 461-478, 2014

work page 2014

[5] [5]

A comprehen- sive review on microgrid and virtual power plant concepts employed for distributed energy resources scheduling in power systems

S. M. Nosratabadi and R.-A. Hooshmand, E. Gholipour, “A comprehen- sive review on microgrid and virtual power plant concepts employed for distributed energy resources scheduling in power systems” Renewable and Sustainable Energy Reviews, vol. 67, pp. 341–363, 2017. a

work page 2017

[6] [6]

Distributed Optimal Power Management via Negawatt Trading in Real-Time Electricity Market

Y . Okawa and T. Namerikawa, “Distributed Optimal Power Management via Negawatt Trading in Real-Time Electricity Market”, IEEE Trans. Smart Grid, vol. 8, pp. 3009-3019, 2017

work page 2017

[7] [7]

Challenges and prospects for negawatt trading in light of recent technological developments

W. Tushar, T. K. Saha, C. Yuen, D. Smith, P. Ashworth, H. V . Poor and S. Basnet, “Challenges and prospects for negawatt trading in light of recent technological developments”, Nat. energy, vol. 5, pp. 834-841, 2020

work page 2020

[8] [8]

A design method for electric power demand portfolio and its basic investigation -Application to design of aggregator portfolio-

T. Tsurumi and K. Tokoro, “A design method for electric power demand portfolio and its basic investigation -Application to design of aggregator portfolio-”, CRIEPI report, C17016 (2018) (in Japanese)

work page 2018