Pith. sign in

REVIEW 4 major objections 5 minor 29 references

QCOEM: Quantum Cloud Orchestration with Evolutionary Multi-Objective Optimization

T0 review · 4 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read QCOEM claims that scheduling quantum tasks by evolutionary multi-objective search—optimizing completion time, error rate, and load imbalance together—yields zero task rescheduling and about 30% higher mean fidelity than noise-agnostic heuri

desk verdict A coherent, re-implementable MOEA scheduling study whose headline fidelity gain is the optimizer improving its own surrogate—useful incremental work, but the central claim needs an independent measurement. read the letter →

arxiv 2607.25358 v1 pith:LDIWKEUT submitted 2026-07-28 cs.DC cs.ETquant-ph

Tam N. Pham , Hoa T. Nguyen , Quan Le-Trung This is my paper
classification cs.DCcs.ETquant-ph
keywords quantumcloudcomputingtaskschedulingmulti-objectiveoptimizationevolutionaryalgorithmNSGA-IINSGA-IIIachievementscalarizationfunctionnoise-awareorchestration
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 tries to establish that quantum cloud orchestration should be treated as a multi-objective optimization problem—balancing completion time, execution error, and load imbalance—rather than a throughput problem. It proposes QCOEM, a scheduler that uses NSGA-II and NSGA-III evolutionary algorithms to approximate the Pareto front of task-to-backend assignments, then applies an Augmented Achievement Scalarization Function to pick a single dispatchable schedule aligned with user priorities. In an emulated heterogeneous cloud with five calibrated noisy backends, the framework reports zero task rescheduling and roughly 30% higher mean fidelity than greedy, round-robin, and random baselines, with bounded scheduling overhead. The underlying claim is that noise-aware, preference-driven routing is practical enough to deploy as a lightweight control-plane service.

What carries the argument

NSGA-II and NSGA-III (non-dominated sorting genetic algorithms) approximate the Pareto-optimal set of task-to-backend assignments across three objectives: mean completion time (Eq. 4), mean error rate (Eq. 6), and load imbalance (Eq. 7). The Augmented Achievement Scalarization Function (Eq. 13) is the preference-based decision rule that selects one schedule from the non-convex front by minimizing a weighted max-plus-epsilon expression, avoiding the duality gap of weighted-sum scalarization. The fidelity model F(τ_i,q_j)=∏(1−ε_{g,j}) carries the argument: because per-gate errors multiply, deeper circuits on noisy backends are penalized, which is what makes routing decisions matter.

What would settle it

Run the QCOEM scheduler against a real noisy backend with known calibration drift and compare realized fidelity (measured by, e.g., state tomography or expectation values) with the product-of-errors prediction. If realized fidelity does not track the surrogate, or if the scheduler's assignments are no better than noise-agnostic baselines once drift is present, the central claim fails. A simpler check is to re-run the emulation under a correlated-error noise model or time-varying calibration snapshots and see whether the ~30% fidelity gap and zero-rescheduling result survive.

Watch

Extended reading notes

Core claim

The paper's central claim is that a scheduler which simultaneously minimizes mean completion time, mean error rate, and load imbalance—then selects a final schedule with an augmented achievement scalarization function—can deliver stable, high-fidelity execution in a heterogeneous quantum cloud. QCOEM encodes each schedule as a binary task-to-QNode assignment, estimates completion time from critical-path gate latencies and heartbeat-based queueing delays, estimates fidelity as the product of per-operation success probabilities from calibrated snapshots, and uses NSGA-II/NSGA-III to evolve a non-dominated set of candidate assignments. The AASF step maps that Pareto set to one dispatchable sche

Load-bearing premise

The reported fidelity gains rest entirely on the model F=∏(1−ε), a product of static per-gate and readout success probabilities drawn from calibration snapshots, which the paper assumes remain stable across runs and never validates against actual noisy execution outcomes or real hardware.

Editorial extensions

If this is right

  • If QCOEM's claims hold, noise-aware multi-objective scheduling can be embedded in a cloud-native control plane without modifying the quantum hardware itself.
  • Preference weights give operators a practical lever: speed-focused, quality-focused, and load-focused policies produce materially different schedules with predictable latency–fidelity trade-offs.
  • Zero rescheduling implies that heartbeat-derived queue estimates and capacity constraints at dispatch time are sufficient to avoid infeasible assignments under stable calibration conditions.
  • The roughly 30% fidelity gain over greedy, round-robin, and random baselines suggests that ignoring backend error rates systematically sacrifices execution quality even when latency is acceptable.
  • Scheduling overhead remains bounded (about 12× greedy at batch size 50), making evolutionary search feasible in settings where QPU queueing dominates end-to-end latency.

Reading between the lines

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

  • A direct testable extension would be to replace the static snapshot fidelity model with per-run calibration data or noise-adaptive transpilation, then re-measure whether the scheduler's routing decisions still improve realized fidelity under drift.
  • The zero-rescheduling result likely depends on calibration snapshots staying stable during a scheduling round; on real backends with calibration drift, QCOEM would need to re-optimize in-flight assignments, and this claim would weaken.
  • The same MOEA-plus-AASF machinery could transfer to other resource-allocation problems with non-convex Pareto fronts and drifting performance, such as serverless function placement over heterogeneous and time-varying infrastructure.
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

4 major / 5 minor

Summary. The paper proposes QCOEM, a quantum cloud orchestration framework that formulates task-to-QNode assignment as a three-objective optimization problem (mean completion time, mean error rate, load imbalance). It uses NSGA-II and NSGA-III to approximate the Pareto front and an Augmented Achievement Scalarization Function (AASF) to select a single dispatchable schedule from user preference weights. The system is implemented on Kubernetes with Qiskit FakeBackendV2 calibration snapshots as QNodes and evaluated against Greedy, Round-Robin, and Random baselines on MQTBench workloads. The headline claims are zero task rescheduling and approximately 30% higher mean fidelity than noise-agnostic heuristics, with bounded scheduling overhead.

Significance. If the fidelity result were independently validated, QCOEM would be a useful contribution: it gives a clean MOOP formulation for quantum scheduling, a reproducible simulation environment (pymoo + Qiskit + MQTBench), a preference-based decision rule (AASF) that handles non-convex Pareto fronts, and an explicit load-imbalance objective, which is missing from several prior schedulers including Qonductor and QFOR. The paper is transparent about hyperparameters, workload construction, and implementation details, and the latency/load results plausibly follow from the stated dynamics. However, the central fidelity claim is not independently established because the reported metric is the same surrogate that the optimizer minimizes, and the 'bounded overhead' claim is contradicted by the largest batch measurement. These issues are fixable but currently make the paper's main quantitative contribution conditional.

major comments (4)
  1. [IV.C, Eq. (5)-(6)] The 'Mean Fidelity' metric is defined as the mean of F(tau_i, q_j) from Eq. (5), and f2 in Eq. (6) is exactly 1 - F(tau_i, q_j). Thus the reported ~30% fidelity gain over noise-agnostic baselines is the optimizer improving its own objective, not an independent quality measure. The paper never compares F to actual noisy execution outcomes: although the QNodes execute circuits with FakeBackendV2 noise models, no output-distribution fidelities are reported, and §VI states real-hardware validation is future work. To support the abstract and conclusion claims of 'stable, high-fidelity execution', the authors should compute achieved fidelities from the simulated execution results (e.g., Hellinger fidelity or sample-based success probability) and show that F predicts them, or explicitly weaken the claim to a surrogate-objective improvement.
  2. [IV.D.3] At batch size M=50, NSGA-II and NSGA-III require 94.58 s and 84.80 s of scheduling overhead, respectively. With the stated Poisson arrival rate lambda=2.0 tasks/s, a batch of 50 tasks forms over roughly 25 s on average. The scheduler cannot keep up with the task arrival stream at this batch size, so the abstract's 'bounded scheduling overhead' is not supported by the reported data. The paper should report overhead against the available scheduling interval (batch interval) and state the conditions under which the MOEA overhead is acceptable, rather than merely listing wall-clock times.
  3. [IV.A.3, IV.D] Section IV.A.3 states that results are reported across multiple random seeds, but no error bars, confidence intervals, or per-seed values appear in Fig. 2 or Fig. 3. In particular, §IV.D.2 claims NSGA-II shows 'lower variance than NSGA-III' and the abstract claims 'stable' execution. Such claims are not assessable without distributional information. Please provide mean ± standard deviation (or equivalent) for all five evaluation metrics, or show per-seed scatter.
  4. [IV.D.2, Fig. 2(b)] The 'zero task rescheduling' result is presented as a headline advantage, but rescheduling rate is defined only as infeasibility at dispatch (e.g., capacity violations or stale state). Since the MOEAs explicitly enforce constraints C2 and C3 while the baselines ignore them, zero rescheduling for the MOEAs is a direct consequence of the constraint formulation. The experiments do not inject QNode failures or calibration drift during a batch, so this result does not substantiate the fault-tolerance discussion in §III.B.4. The claim should be scoped to constraint satisfaction, or the experiments should include dynamic failure/drift scenarios.
minor comments (5)
  1. [II.B, Eq. (13)] The ideal point z* is not formally defined; it should be stated as the vector of per-objective minima (or the normalized ideal point). Also, Eq. (13) divides by w_k, so the text should explicitly require w_k > 0 for all k (the experiments use 0.05, but the general formulation does not say this).
  2. [IV.D.3] The Greedy scheduling time of 0.78 s for M=5 seems surprisingly high for a heuristic that should be near-instant; please clarify whether this includes transpilation or setup costs. If so, state what is excluded from the MOEA measurements.
  3. [IV.D] Terms such as 'total completion time', 'MCT', and 'execution time' are used somewhat interchangeably. In particular, §IV.D.2(a) uses 'total completion time' while Fig. 3(a) says 'Mean task completion time'; please standardize notation and define each quantity in one place.
  4. [Fig. 2 and Fig. 3] Captions do not specify whether bars show means, medians, or single runs; adding units and statistical summaries would improve interpretability.
  5. [V] Qonductor and QFOR are cited as noise-aware schedulers, but no quantitative comparison is made with them. While the paper's baseline choice is defensible, adding a discussion of expected relative behavior would help position QCOEM against these state-of-the-art methods.

Circularity Check

1 steps flagged · score 6.0 of 10

The 30% fidelity claim is the complement of the optimized objective F, so it is true by construction.

  1. self definitional [Sec. II.A.2 (Eqs. 5–6); Sec. IV.C (Evaluation Metrics)]
    ""Mean Fidelity: the mean of the estimated fidelity F(τi, qj) (Section II), reported as a counterpart of the error-rate objective f2." — Sec. IV.C. "f2(X) = 1/M ∑_{i=1}^M ∑_{j=1}^N x_{ij} · (1− F(τi, qj))" — Eq. (6). The corresponding error rate per-task is 1−F(τi, qj)."

    The metric called 'Mean Fidelity' is the arithmetic mean of F(τi,qj), while objective f2 is the mean of 1−F(τi,qj). Since every task is assigned to exactly one node (C2), Mean Fidelity ≡ 1 − f2. Therefore the reported ~30% fidelity gain over noise-agnostic heuristics is just the statement that the MOEA, whose AASF weights include w_fid=0.45, attains lower values of its own f2 objective than heuristics that ignore that objective. No independent execution-quality measurement is made; the paper's future-work section explicitly defers validation on real hardware. The fidelity result is thus true by construction of the metric/objective pair.

full rationale

The central fidelity claim reduces to the optimizer's own objective. Eq. (5) defines F as the product of static per-gate/readout success probabilities; Eq. (6) minimizes the mean of 1−F; and Sec. IV.C defines 'Mean Fidelity' as the mean of F. Hence comparing QCOEM to noise-agnostic baselines on mean fidelity is comparing values of the very objective that NSGA-II/NSGA-III are designed to minimize. No evidence is provided that F predicts actual noisy-execution fidelity; the paper states snapshots are assumed stable and defers real-hardware validation to future work. The latency, load-imbalance, and overhead results are more independent of this definitional pairing. No load-bearing self-citation chain is present (the cited prior work by the same authors is not used to force the framework's core choices). Therefore the circularity is real but partial, confined to the headline fidelity metric: score 6.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The framework relies on a small set of modeling assumptions (surrogate fidelity, critical-path timing, heartbeat queue estimates) plus standard MOEA/AASF machinery. There are no new physical entities or fitted constants beyond preference weights and algorithm hyperparameters; the main unvalidated input is the product-of-success fidelity surrogate.

free parameters (5)
  • User preference weight vector w = Balanced [0.35,0.45,0.20]; Speed-Focus [0.90,0.05,0.05]; Quality-Focus [0.05,0.90,0.05]; Load-Focus [0.05,0.05,0.90]
    AASF selection depends on these weights; reported MCT, fidelity, and load-imbalance results change materially across the four settings.
  • AASF augmentation parameter ρ = 1e-4
    Chosen to avoid degenerate scalarization; standard small constant but still a hand-set value affecting selected schedule.
  • MOEA population size P and max generations Gmax = P=300, Gmax=1000
    Algorithm hyperparameters that directly affect solution quality and the reported ~95s scheduling overhead at M=50.
  • Task arrival rate λ = 2.0 tasks/s
    Poisson arrival rate drives the queueing dynamics and load-imbalance values; chosen by the authors for the workload study.
  • Batch size M = 5, 10, 20, 50
    Scaling experiments vary batch size; overhead grows with M and is a key factor in the 'bounded overhead' claim.
assumptions (4)
  • domain assumption Task fidelity is the product of independent per-operation success probabilities (Eq. 5); errors are uncorrelated and calibration snapshots are stable during a scheduling round.
    Underlies objective f2 and the reported mean-fidelity metric; no validation against real hardware or correlated noise models.
  • domain assumption Execution time is critical-path gate latency times number of shots, with gates and shots executed sequentially (Eq. 1).
    Used to compute f1 and queue updates; the model is simplistic and may misestimate real backend execution time.
  • domain assumption QNode heartbeat next-available time and queue state remain accurate until dispatch; no arrivals occur during the optimization round.
    The scheduler optimizes using cached H_j; with ~95s optimization times, queue states can become stale and degrade real-world performance.
  • domain assumption FakeBackendV2 calibration snapshots represent real IBM backend behavior.
    All fidelity and timing inputs come from emulated snapshots, not live devices; real-hardware drift and correlated noise are not captured.

how reviews work

0 comments
Cite this review

Pith. "Pith review of QCOEM: Quantum Cloud Orchestration with Evolutionary Multi-Objective Optimization." pith.science (2026). https://pith.science/paper/LDIWKEUT

@misc{pith2026260725358,
  author       = {Pith},
  title        = {Pith review of: QCOEM: Quantum Cloud Orchestration with Evolutionary Multi-Objective Optimization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/LDIWKEUT}},
  note         = {Machine review of arXiv:2607.25358}
}
read the original abstract

Quantum cloud platforms need to dynamically orchestrate workloads across heterogeneous quantum computation backends whose noise profiles, qubit topologies, and queues vary over time. Existing orchestrators use noise-agnostic heuristics that ignore backend-specific errors, causing reduced execution fidelity, load imbalance, and frequent rescheduling. To address these challenges, we propose QCOEM - a Quantum Cloud Orchestration framework that leverages Evolutionary algorithms for Multi-objective optimization of quantum task scheduling. We compare NSGA-II and NSGA-III for jointly minimizing mean completion time, execution error rate, and load imbalance. To select schedules from a non-convex Pareto front, we apply an Augmented Achievement Scalarization Function (AASF) as a preference-based decision rule that maps the Pareto set to a single dispatchable schedule aligned with user priorities. Our extensive performance evaluation in a heterogeneous quantum cloud environment shows zero task rescheduling and about 30% higher mean fidelity than noise-agnostic heuristics, while maintaining bounded scheduling overhead. The experiment results indicate that our QCOEM framework can deliver stable, high-fidelity execution and lightweight resource management for quantum cloud computing.

Figures

Figures reproduced from arXiv: 2607.25358 by the authors.

Figure 1
Figure 1. System overview of the proposed QCOEM framework. Periodic [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Performance evaluation with baseline approaches: (a) time efficiency [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Performance under different preference settings (bal: Balance, speed: [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

29 extracted references · 2 linked inside Pith

  1. [1]

    Adaptive job and resource management for the growing quantum cloud,

    G. S. Ravi, K. N. Smith, P. Murali, and F. T. Chong, “Adaptive job and resource management for the growing quantum cloud,” in2021 IEEE International Conference on Quantum Computing and Engineering (QCE). IEEE, 2021, pp. 301–312

    2021

  2. [2]

    Quantum Cloud Computing: A Review, Open Problems, and Future Directions,

    H. T. Nguyen, P. Krishnan, D. Krishnaswamy, M. Usman, and R. Buyya, “Quantum Cloud Computing: A Review, Open Problems, and Future Directions,” Apr. 2024, arXiv:2404.11420

    arXiv 2024

  3. [3]

    Quantum computing in the nisq era and beyond,

    J. Preskill, “Quantum computing in the nisq era and beyond,”Quantum, vol. 2, p. 79, 2018

    2018

  4. [4]

    Qonductor: A cloud orchestrator for quantum computing,

    E. Giortamis, F. Romao, N. Tornow, D. Lugovoy, and P. Bhatotia, “Qonductor: A cloud orchestrator for quantum computing,” inProceed- ings of the International Conference for High Performance Computing, Networking, Storage and Analysis, ser. SC ’25. New York, NY , USA: ACM, 2025, p. 728–745

    2025

  5. [5]

    Qubernetes: Towards a unified cloud-native execution platform for hybrid classic- quantum computing,

    V . Stirbu, O. Kinanen, M. Haghparast, and T. Mikkonen, “Qubernetes: Towards a unified cloud-native execution platform for hybrid classic- quantum computing,”Information and Software Technology, vol. 175, p. 107529, 2024

    2024

  6. [6]

    Middleware for quantum: An orchestration of hybrid quantum- classical systems,

    I. Faro, I. Sitdikov, D. G. Valiñas, F. J. M. Fernandez, C. Codella, and J. Glick, “Middleware for quantum: An orchestration of hybrid quantum- classical systems,” in2023 IEEE International Conference on Quantum Software (QSW). IEEE, 2023, pp. 1–8

    2023

  7. [7]

    Qusplit: Achieving both high fidelity and throughput via job splitting on noisy quantum computers,

    J. Li, Y . Song, Y . Liu, J. Pan, L. Yang, T. Humble, and W. Jiang, “Qusplit: Achieving both high fidelity and throughput via job splitting on noisy quantum computers,”arXiv preprint arXiv:2501.12492, 2025

    arXiv 2025

  8. [8]

    Performance comparison of heuristic algorithms for task scheduling in iaas cloud computing environment,

    S. H. H. Madni, M. S. Abd Latiff, M. Abdullahi, S. M. Abdulhamid, and M. J. Usman, “Performance comparison of heuristic algorithms for task scheduling in iaas cloud computing environment,”PloS one, vol. 12, no. 5, p. e0176321, 2017

    2017

Show all 29 references
  1. [9]

    Heuristic algorithms for task scheduling in cloud computing: a survey,

    N. Soltani, B. Soleimani, and B. Barekatain, “Heuristic algorithms for task scheduling in cloud computing: a survey,”International Journal of Computer Network and Information Security, vol. 9, no. 8, p. 16, 2017

    2017

  2. [10]

    Qos: quantum operating system,

    E. Giortamis, F. Romão, N. Tornow, and P. Bhatotia, “Qos: quantum operating system,” inProceedings of the 19th USENIX Conference on Operating Systems Design and Implementation, ser. OSDI ’25. USA: USENIX Association, 2025

    2025

  3. [11]

    Optimized noise suppression for quantum circuits,

    F. Wagner, D. J. Egger, and F. Liers, “Optimized noise suppression for quantum circuits,”INFORMS J. on Computing, vol. 37, no. 1, p. 22–41, Jan. 2025

    2025

  4. [12]

    Noise-adaptive compiler mappings for noisy intermediate-scale quan- tum computers,

    P. Murali, J. M. Baker, A. Javadi-Abhari, F. T. Chong, and M. Martonosi, “Noise-adaptive compiler mappings for noisy intermediate-scale quan- tum computers,” inProceedings of the Twenty-Fourth International Conference on Architectural Support for Programming Languages and Oper...

    2019

  5. [13]

    An evolutionary many-objective optimization algorithm using reference-point-based nondominated sorting approach, part i: solving problems with box constraints,

    K. Deb and H. Jain, “An evolutionary many-objective optimization algorithm using reference-point-based nondominated sorting approach, part i: solving problems with box constraints,”IEEE transactions on evolutionary computation, vol. 18, no. 4, pp. 577–601, 2013

    2013

  6. [14]

    QFOR: A Fidelity-aware Orchestrator for Quantum Computing Environments using Deep Rein- forcement Learning,

    H. T. Nguyen, M. Usman, and R. Buyya, “QFOR: A Fidelity-aware Orchestrator for Quantum Computing Environments using Deep Rein- forcement Learning,”ACM Transactions on Quantum Computing, Mar. 2026

    2026

  7. [15]

    A fast and elitist multiobjective genetic algorithm: Nsga-ii,

    K. Deb, A. Pratap, S. Agarwal, and T. Meyarivan, “A fast and elitist multiobjective genetic algorithm: Nsga-ii,”IEEE transactions on evolu- tionary computation, vol. 6, no. 2, pp. 182–197, 2002

    2002

  8. [16]

    Jain and K

    H. Jain and K. Deb, “An evolutionary many-objective optimization algorithm using reference-point based nondominated sorting approach, part ii: Handling constraints and extending to an adaptive approach,” IEEE Transactions on evolutionary computation, vol. 18, no. 4, pp. 602– 622, 2013

    2013

  9. [17]

    Multi-objective optimisation using evolutionary algorithms: an introduction,

    K. Deb, “Multi-objective optimisation using evolutionary algorithms: an introduction,” inMulti-objective evolutionary optimisation for product design and manufacturing. Springer, 2011, pp. 3–34

    2011

  10. [18]

    A mathematical basis for satisficing decision making,

    A. P. Wierzbicki, “A mathematical basis for satisficing decision making,” Mathematical modelling, vol. 3, no. 5, pp. 391–405, 1982

    1982

  11. [19]

    Investigating the equivalence between pbi and aasf scalarization for multi-objective optimization,

    H. K. Singh and K. Deb, “Investigating the equivalence between pbi and aasf scalarization for multi-objective optimization,”Swarm and Evolutionary Computation, vol. 53, p. 100630, 2020

    2020

  12. [20]

    pymoo: Multi-objective optimization in python,

    J. Blank and K. Deb, “pymoo: Multi-objective optimization in python,” IEEE Access, vol. 8, pp. 89 497–89 509, 2020

    2020

  13. [21]

    MQT Bench: Bench- marking software and design automation tools for quantum computing,

    N. Quetschlich, L. Burgholzer, and R. Wille, “MQT Bench: Bench- marking software and design automation tools for quantum computing,” Quantum, 2023, MQT Bench is available at https://www.cda.cit.tum.de/ mqtbench/

    2023

  14. [22]

    Drlq: A deep reinforce- ment learning-based task placement for quantum cloud computing,

    H. T. Nguyen, M. Usman, and R. Buyya, “Drlq: A deep reinforce- ment learning-based task placement for quantum cloud computing,” inProceedings of the IEEE 17th International Conference on Cloud Computing (CLOUD 2024). IEEE, 2024, pp. 475–481

    2024

  15. [23]

    Moirai: Optimizing quantum serverless function or- chestration via device allocation and circuit deployment,

    T. Li and Z. Zhao, “Moirai: Optimizing quantum serverless function or- chestration via device allocation and circuit deployment,” inProceedings of the 2024 IEEE International Conference on Web Services (ICWS). IEEE, 2024, pp. 707–717

    2024

  16. [24]

    A conceptual architecture for a quantum-hpc middleware,

    N. Saurabh, S. Jha, and A. Luckow, “A conceptual architecture for a quantum-hpc middleware,” in2023 IEEE international conference on quantum software (QSW). IEEE, 2023, pp. 116–127

    2023

  17. [25]

    iQuantum: a toolkit for mod- eling and simulation of quantum computing environments,

    H. T. Nguyen, M. Usman, and R. Buyya, “iQuantum: a toolkit for mod- eling and simulation of quantum computing environments,”Software: Practice and Experience, vol. 54, no. 6, pp. 1141–1171, Jun. 2024

    2024

  18. [26]

    A digital twin of scalable quantum clouds,

    W. Luo, B. Baheri, T. Humble, J. Zhao, T. Zhan, R. Maharjan, and Q. Guan, “A digital twin of scalable quantum clouds,” inProceedings of the 39th ACM SIGSIM Conference on Principles of Advanced Discrete Simulation, ser. SIGSIM-PADS ’25. New York, NY , USA: ACM, 2025, p. 165–175

    2025

  19. [27]

    Qucloudsim: A customizable discrete event simulator for quantum cloud computing enviroment,

    R. Zhou, Y . Gan, L. Shen, Y . Liu, M. Messel, and C. Qian, “Qucloudsim: A customizable discrete event simulator for quantum cloud computing enviroment,” in2025 International Conference on Quantum Communi- cations, Networking, and Computing (QCNC), 2025, pp. 58–65

    2025

  20. [28]

    Qfaas: A serverless function- as-a-service framework for quantum computing,

    H. T. Nguyen, M. Usman, and R. Buyya, “Qfaas: A serverless function- as-a-service framework for quantum computing,”Future Generation Computer Systems, vol. 154, pp. 281–300, 2024

    2024

  21. [29]

    Adaptive job scheduling in quantum clouds using reinforcement learning,

    W. Luo, J. Zhao, T. Zhan, and Q. Guan, “Adaptive job scheduling in quantum clouds using reinforcement learning,” inProceedings of the 54th International Conference on Parallel Processing, ser. ICPP ’25. New York, NY , USA: ACM, 2025, p. 658–667

    2025

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.