pith. sign in

arxiv: 2604.24340 · v1 · submitted 2026-04-27 · 💻 cs.DC

Exact, Efficient, and Reliable Multi-Objective and Multi-Constrained IoT Workflow Scheduling in Edge-Hub-Cloud Cyber-Physical Systems

Andreas Kouloumpris , Georgios L. Stavrinides , Maria K. Michael , Theocharis Theocharides This is my paper

Pith reviewed 2026-05-08 01:41 UTC · model grok-4.3

classification 💻 cs.DC
keywords IoT workflow schedulingedge-hub-cloud systemsmulti-objective optimizationmixed integer linear programmingtask duplicationlatencyenergy efficiencyreliability
0
0 comments X

The pith

A continuous-time mixed integer linear programming formulation optimally schedules IoT workflows across edge, hub, and cloud to jointly minimize latency and energy while maximizing reliability.

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

The paper develops an exact optimization technique for scheduling interdependent tasks in IoT cyber-physical systems that run on edge devices with multicore processors, a hub device, and a cloud server. It uses mixed integer linear programming to simultaneously reduce latency and energy use while increasing reliability, all subject to deadlines, resource limits, and other constraints. Selective task duplication boosts reliability without the overhead of full replication. Tests on a real IoT workflow and synthetic graphs of varying sizes show the exact method beats an extended heuristic by up to 29.83 percent in latency, 33.96 percent in energy, and 28.49 percent in reliability, while keeping computation times practical. This matters for critical applications that require dependable execution on heterogeneous and constrained hardware.

Core claim

We propose a continuous-time mixed integer linear programming formulation for multi-objective and multi-constrained workflow scheduling in edge-hub-cloud cyber-physical systems. The formulation jointly optimizes latency, energy, and reliability while holistically addressing timing, capability, memory, storage, and energy constraints. It selectively employs task duplication to enhance reliability without unnecessary overhead. Evaluation on a real-world IoT workflow and synthetic task graphs across different system configurations and objective trade-offs shows average improvements of up to 29.83 percent in latency, 33.96 percent in energy, and 28.49 percent in reliability over a widely used,公平

What carries the argument

Continuous-time mixed integer linear programming formulation that models task assignments to heterogeneous processors, execution timings, and selective duplications to optimize the weighted multi-objective function under all constraints.

If this is right

  • Schedules produced by the formulation simultaneously improve all three objectives for task graphs typical of targeted IoT applications.
  • Practical runtimes enable use in offline planning or periodic recomputation for static or slowly changing workflows.
  • Changing the weights in the objective function lets users explore concrete latency-energy-reliability trade-offs.
  • Selective duplication yields reliability gains while avoiding the full energy and latency cost of replicating every task.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The MILP could serve as a quality benchmark for developing faster approximation or heuristic methods for similar edge-cloud scheduling problems.
  • Extending the model to handle runtime changes in task graphs or available resources would test its usefulness for adaptive IoT applications.
  • Hardware-specific measurement of timing and energy parameters could be used to tighten the model and reduce any gap between predicted and observed performance.

Load-bearing premise

The continuous-time mixed integer linear programming model accurately represents the real-world timing, energy consumption, and reliability behaviors of the heterogeneous multicore processors and system components without significant discrepancies.

What would settle it

Running the generated schedules on a physical edge-hub-cloud hardware testbed, measuring actual end-to-end latency, energy draw, and observed reliability, then checking whether the percentage improvements over the heuristic match the reported figures.

Figures

Figures reproduced from arXiv: 2604.24340 by Andreas Kouloumpris, Georgios L. Stavrinides, Maria K. Michael, Theocharis Theocharides.

Figure 1
Figure 1. Figure 1: Examples of critical IoT-enabled cyber-physical applications following view at source ↗
Figure 2
Figure 2. Figure 2: TG transformation example: (a) considered CPS, (b) TG view at source ↗
Figure 3
Figure 3. Figure 3: Overview of our extension to HEFT. each set of τi (lines 41–44). Finally, we select for τi the set of candidate nodes that minimizes the multi-objective function, without exceeding the defined deadline (lines 45–55), as in (35). If there is a task for which no set satisfies all of the above constraints, then the problem is infeasible. The output of extended HEFT includes the selected candidate nodes and ar… view at source ↗
Figure 4
Figure 4. Figure 4: Examined CPS. TABLE III DEVICE CAPABILITIES ca Capability Device Task Type Type 0 Basic computational capability All All 1 Thermal camera e Entry 2 LiDAR sensor e Entry 3 Multispectral camera e Entry 4 High-precision GNSS1 module e Interm. 5 Tag release mechanism e Exit 6 UAV coordination module h Interm. 7 Integrated display h Exit 8 High-performance GPU c Interm. 9 High-availability storage c Exit 1Globa… view at source ↗
Figure 5
Figure 5. Figure 5: Real-world IoT workflow TG. Li,µk.q, and Pi,µk.q were determined through profiling and power monitoring tools (perf and Powertop) [44] across all system devices. The reliability threshold Rthr i of each task was set within [0.9990, 0.9999], so that based on (9) it would not always be met by solely executing the primary task [41]. The values of the ETAG parameters are listed in Table IX. Ei,µk.q, Ri,µk.q, ζ… view at source ↗
Figure 6
Figure 6. Figure 6: Comparison between proposed MILP approach and extended HEFT view at source ↗
Figure 8
Figure 8. Figure 8: Comparison between proposed MILP approach and extended HEFT view at source ↗
Figure 10
Figure 10. Figure 10: Solver runtime for proposed MILP method under increasing TG view at source ↗
Figure 9
Figure 9. Figure 9: Improvement in latency, energy, and reliability objectives attained by view at source ↗
read the original abstract

Emerging IoT-enabled cyber-physical applications demand low-latency, energy-efficient, and reliable execution across resource-constrained edge devices with heterogeneous multicore processors and diverse sensing and actuating capabilities, in collaboration with a hub device and a cloud server. These workflow-based applications comprise interdependent tasks that must be executed under stringent deadline, reliability, capability, memory, storage, and energy constraints. Given their critical nature, exact optimization is necessary to obtain optimal schedules that ensure dependable operation. Existing scheduling approaches, both exact and heuristic, fail to jointly address all these objectives and constraints. To this end, we propose an exact multi-objective and multi-constrained workflow scheduling approach for edge-hub-cloud cyber-physical systems, based on continuous-time mixed integer linear programming. The proposed formulation jointly optimizes latency, energy, and reliability, while holistically addressing timing and resource constraints. To enhance reliability while avoiding the overhead of unnecessary task replicas, it selectively employs task duplication. We evaluate our approach against a widely used heuristic, which we extend to ensure a fair and meaningful comparison, using a real-world IoT workflow and synthetic task graphs of varying sizes, across different system configurations and objective trade-offs. The proposed method consistently outperforms the heuristic, achieving up to 29.83%, 33.96%, and 28.49% average improvements in latency, energy, and reliability, respectively, while attaining practical runtimes. Overall, the experimental results demonstrate the effectiveness of our approach under various system configurations and objective trade-offs, and show its practical scalability to task graphs of sizes relevant to the targeted applications and system architecture.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper proposes a continuous-time mixed-integer linear programming (MILP) formulation for exact joint optimization of latency, energy, and reliability in scheduling IoT workflows across edge-hub-cloud cyber-physical systems. The approach incorporates multiple constraints (deadlines, reliability, capability, memory, storage, energy) and uses selective task duplication to improve reliability without excessive overhead. It is evaluated on a real-world IoT workflow and synthetic task graphs of varying sizes under different system configurations and objective weightings, claiming consistent outperformance over an extended heuristic baseline with average improvements of 29.83% in latency, 33.96% in energy, and 28.49% in reliability, while maintaining practical runtimes.

Significance. If the continuous-time model accurately captures hardware behavior, the work provides a valuable exact optimization framework for critical multi-objective CPS scheduling problems where heuristics may miss globally optimal trade-offs. The selective duplication mechanism and holistic constraint handling address practical deployment needs in heterogeneous multicore edge environments. The reported practical runtimes for relevant workflow sizes and the use of both real and synthetic inputs are strengths that support potential applicability.

major comments (2)
  1. [§3] §3 (System Model and Problem Formulation): The continuous-time MILP models task execution, power draw, and reliability on heterogeneous multicore processors using idealized linear or piecewise-linear abstractions. This is load-bearing for the central performance claims because the reported improvements (e.g., 29.83% latency) are computed entirely within this model; discrete effects such as cache contention, context-switch costs, DVFS transients, and stochastic inter-device delays are not modeled, risking overestimation of real-world gains.
  2. [§5] §5 (Experimental Evaluation): All comparisons and metric improvements are obtained by solving the MILP and running the heuristic inside the same continuous-time simulator. No hardware-in-the-loop validation or trace-based calibration against measured execution times, energy profiles, or failure rates on the target multicore platforms is presented, so it remains unclear whether the 28–34% gains would hold when the schedules are deployed on actual edge-hub-cloud hardware.
minor comments (2)
  1. [§4] The objective function weighting scheme and the exact definition of the reliability metric (including how duplication affects failure probability) would benefit from an explicit equation reference in the formulation section for reproducibility.
  2. [§5] Figure captions and axis labels in the trade-off plots (e.g., latency vs. energy Pareto fronts) could be expanded to indicate the specific system configuration and objective weights used for each curve.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript. We address each major comment below and will revise the paper to better articulate the scope and limitations of our modeling and evaluation approach.

read point-by-point responses

  1. Referee: [§3] §3 (System Model and Problem Formulation): The continuous-time MILP models task execution, power draw, and reliability on heterogeneous multicore processors using idealized linear or piecewise-linear abstractions. This is load-bearing for the central performance claims because the reported improvements (e.g., 29.83% latency) are computed entirely within this model; discrete effects such as cache contention, context-switch costs, DVFS transients, and stochastic inter-device delays are not modeled, risking overestimation of real-world gains.

    Authors: We agree that the continuous-time MILP relies on linear and piecewise-linear abstractions for task execution, power draw, and reliability. These choices enable an exact, tractable formulation while capturing the primary timing, energy, and reliability dynamics of heterogeneous multicore edge-hub-cloud systems. The reported improvements are measured relative to this model. In the revision we will insert a dedicated subsection in §3 that explicitly lists the modeling assumptions, discusses the omitted effects (cache contention, DVFS transients, context-switch overhead, and stochastic delays), and qualifies the conditions under which the predicted gains are expected to translate to hardware. This addition will not alter the core technical contribution but will improve interpretability. revision: yes

  2. Referee: [§5] §5 (Experimental Evaluation): All comparisons and metric improvements are obtained by solving the MILP and running the heuristic inside the same continuous-time simulator. No hardware-in-the-loop validation or trace-based calibration against measured execution times, energy profiles, or failure rates on the target multicore platforms is presented, so it remains unclear whether the 28–34% gains would hold when the schedules are deployed on actual edge-hub-cloud hardware.

    Authors: We acknowledge that every quantitative result is generated inside the continuous-time simulator and that no hardware-in-the-loop or trace-driven calibration on physical multicore platforms is provided. This is a standard limitation when introducing a new exact multi-objective MILP formulation, because obtaining consistent, fine-grained measurements across all modeled parameters on the target heterogeneous hardware is a separate, resource-intensive effort. The real-world IoT workflow supplies the task-graph topology and constraint values, yet we agree that direct deployment data would strengthen confidence in the gains. In the revised manuscript we will expand §5 and the conclusions to (i) state clearly that results are model-based, (ii) reference any publicly available platform traces that informed our parameter ranges, and (iii) outline concrete steps for future hardware validation. No new experimental data will be added in this revision. revision: partial

Circularity Check

0 steps flagged

No circularity: standard MILP formulation with independent empirical evaluation

full rationale

The paper introduces a continuous-time MILP model that directly encodes the scheduling objectives (latency, energy, reliability) and constraints (deadlines, resources, selective duplication) as linear inequalities and objective functions. The claimed performance gains are produced by solving this model to optimality and comparing the resulting schedules against an extended heuristic baseline on both a real IoT workflow and synthetic DAGs. No step reduces a derived quantity to a fitted parameter by construction, no load-bearing claim rests on a self-citation chain, and the formulation is not an ansatz smuggled from prior author work. The derivation is therefore self-contained and externally falsifiable via the reported runtime and improvement numbers.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Based on the abstract only, no specific free parameters, axioms, or invented entities are detailed. The approach relies on standard MILP modeling assumptions for scheduling problems.

pith-pipeline@v0.9.0 · 5617 in / 1030 out tokens · 51038 ms · 2026-05-08T01:41:11.482544+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

48 extracted references · 1 canonical work pages

  1. [1]

    Resource scheduling and delay optimiza- tion of IoT devices in drone-assisted multiaccess edge computing,

    L. Qu, J. Wang, and C. Assi, “Resource scheduling and delay optimiza- tion of IoT devices in drone-assisted multiaccess edge computing,”IEEE Internet Things J., vol. 12, no. 11, pp. 16 998–17 011, Jun. 2025

    2025

  2. [2]

    IoT systems for extreme environments,

    K. Kant, A. Jolfaei, and K. Moessner, “IoT systems for extreme environments,”IEEE Internet Things J., vol. 11, no. 3, pp. 3671–3675, Feb. 2024

    2024

  3. [3]

    An optimization framework for task allocation in the edge/hub/cloud paradigm,

    A. Kouloumpris, G. L. Stavrinides, M. K. Michael, and T. Theocharides, “An optimization framework for task allocation in the edge/hub/cloud paradigm,”Future Gener. Comput. Syst., vol. 155, pp. 354–366, Jun. 2024

    2024

  4. [4]

    Energy and cost-aware workflow scheduling in cloud computing data centers using a multi-objective optimization algorithm,

    A. Mohammadzadeh, M. Masdari, and F. S. Gharehchopogh, “Energy and cost-aware workflow scheduling in cloud computing data centers using a multi-objective optimization algorithm,”J. Netw. Syst. Manag., vol. 29, no. 3, pp. 31–64, Apr. 2021

    2021

  5. [5]

    Self-powered cardiovascular electronic devices and systems,

    Q. Zheng, Q. Tang, Z. L. Wang, and Z. Li, “Self-powered cardiovascular electronic devices and systems,”Nat. Rev. Cardiol., vol. 18, no. 1, pp. 7–21, Jan. 2021

    2021

  6. [6]

    BESI: Reliable and heterogeneous sensing and intervention for in-home health applications,

    R. Alam, J. Dugan, N. Homdee, N. Gandhi, B. Ghaemmaghami, H. Meda, A. Bankole, M. Anderson, J. Gong, T. Smith-Jackson, and J. Lach, “BESI: Reliable and heterogeneous sensing and intervention for in-home health applications,” inProc. IEEE/ACM Int. Conf. Connect. Health Appl. Syst. Eng. Technol. (CHASE), 2017, pp. 147–156

    2017

  7. [7]

    Multi-UA V based autonomous wilderness search and rescue using target iso-probability curves,

    Z. Kashino, G. Nejat, and B. Benhabib, “Multi-UA V based autonomous wilderness search and rescue using target iso-probability curves,” in Proc. Int. Conf. Unmanned Aircr. Syst. (ICUAS), 2019, pp. 636–643

    2019

  8. [8]

    ICARUS: Automatic autonomous power infrastructure inspection with UA Vs,

    A. Savva, A. Zacharia, R. Makrigiorgis, A. Anastasiou, C. Kyrkou, P. Kolios, C. Panayiotou, and T. Theocharides, “ICARUS: Automatic autonomous power infrastructure inspection with UA Vs,” inProc. Int. Conf. Unmanned Aircr. Syst. (ICUAS), 2021, pp. 918–926

    2021

  9. [9]

    Resource management for mission-critical applications in edge computing: Systematic review on recent research and open issues,

    N. Rasouli, C. Klein, and E. Elmroth, “Resource management for mission-critical applications in edge computing: Systematic review on recent research and open issues,”ACM Comput. Surv., vol. 58, no. 3, pp. 71:1–71:37, Sep. 2025

    2025

  10. [10]

    SDN-based NFV deployment for multi-objective resource allocation in edge com- puting: A deep reinforcement learning for IoT workload scheduling,

    M. Hosseinzadeh, A. Haider, A. M. Rahmani, F. S. Gharehchopogh, S. Rajabi, P. Khoshvaght, T. Porntaveetus, and S.-W. Lee, “SDN-based NFV deployment for multi-objective resource allocation in edge com- puting: A deep reinforcement learning for IoT workload scheduling,” Sustain. Comput. Inform. Syst., vol. 48, p. 101218, Dec. 2025

    2025

  11. [11]

    A self-supervised deep reinforcement learning for zero-shot task scheduling in mobile edge computing environments,

    P. Khoshvaght, A. Haider, A. M. Rahmani, S. Rajabi, F. S. Ghare- hchopogh, J. Lansky, and M. Hosseinzadeh, “A self-supervised deep reinforcement learning for zero-shot task scheduling in mobile edge computing environments,”Ad Hoc Netw., vol. 178, p. 103977, Nov. 2025

    2025

  12. [12]

    Dependable multicore architectures at nanoscale: The view from Europe,

    M. Ottavi, S. Pontarelli, D. Gizopoulos, C. Bolchini, M. K. Michael, L. Anghel, M. Tahoori, A. Paschalis, P. Reviriego, and O. e. a. Bring- mann, “Dependable multicore architectures at nanoscale: The view from Europe,”IEEE Des. Test, vol. 32, no. 2, pp. 17–28, Apr. 2015

    2015

  13. [13]

    Binary search-based fast scheduling algorithms for reliability-aware energy-efficient task graph scheduling with fault tolerance,

    S. K. Biswas, P. K. Muhuri, and U. K. Roy, “Binary search-based fast scheduling algorithms for reliability-aware energy-efficient task graph scheduling with fault tolerance,”IEEE Trans. Sustain. Comput., vol. 9, no. 3, pp. 433–451, May 2024

    2024

  14. [14]

    Revisiting vulnerability analysis in modern microprocessors,

    M. Maniatakos, M. Michael, C. Tirumurti, and Y . Makris, “Revisiting vulnerability analysis in modern microprocessors,”IEEE Trans. Com- put., vol. 64, no. 9, pp. 2664–2674, Sep. 2015

    2015

  15. [15]

    Normalization and other topics in multi-objective optimization,

    O. Grodzevich and O. Romanko, “Normalization and other topics in multi-objective optimization,” inProc. First Fields–MITACS Ind. Probl. Workshop, 2006, pp. 89–101

    2006

  16. [16]

    Mixed integer linear programming in process scheduling: Modeling, algorithms, and applications,

    C. A. Floudas and X. Lin, “Mixed integer linear programming in process scheduling: Modeling, algorithms, and applications,”Ann. Oper. Res., vol. 139, no. 1, pp. 131–162, Oct. 2005

    2005

  17. [17]

    Optimal multi-constrained workflow scheduling for cyber-physical sys- tems in the edge-cloud continuum,

    A. Kouloumpris, G. L. Stavrinides, M. K. Michael, and T. Theocharides, “Optimal multi-constrained workflow scheduling for cyber-physical sys- tems in the edge-cloud continuum,” inProc. IEEE Annu. Comput. Softw. Appl. Conf. (COMPSAC), 2024, pp. 483–492

    2024

  18. [18]

    Performance-effective and low- complexity task scheduling for heterogeneous computing,

    H. Topcuoglu, S. Hariri, and M.-Y . Wu, “Performance-effective and low- complexity task scheduling for heterogeneous computing,”IEEE Trans. Parallel Distrib. Syst., vol. 13, no. 3, pp. 260–274, Mar. 2002

    2002

  19. [19]

    A functional programming model for embedded dataflow applications,

    C. K ¨uhbacher, C. Mellwig, F. Haas, and T. Ungerer, “A functional programming model for embedded dataflow applications,” inProc. IEEE Annu. Comput. Softw. Appl. Conf. (COMPSAC), 2019, pp. 646–651

    2019

  20. [20]

    On the task mapping and scheduling for DAG-based embedded vision applications on heteroge- neous multi/many-core architectures,

    S. Aldegheri, N. Bombieri, and H. Patel, “On the task mapping and scheduling for DAG-based embedded vision applications on heteroge- neous multi/many-core architectures,” inProc. Des. Autom. Test Eur. Conf. Exhib. (DATE), 2020, pp. 1003–1006

    2020

  21. [21]

    Time- discretization for speeding-up scheduling of deadline-constrained work- flows in clouds,

    T. A. L. Genez, L. F. Bittencourt, and E. R. M. Madeira, “Time- discretization for speeding-up scheduling of deadline-constrained work- flows in clouds,”Future Gener. Comput. Syst., vol. 107, pp. 1116–1129, Jun. 2020

    2020

  22. [22]

    Adaptive real-time scheduling for cooperative cyber-physical systems,

    G. v. Zengen, J. Yu, and L. C. Wolf, “Adaptive real-time scheduling for cooperative cyber-physical systems,” inProc. IEEE Conf. Ind. Cyberphys. Syst. (ICPS), 2020, pp. 79–86

    2020

  23. [23]

    Energy-efficient application mapping and scheduling for lifetime guaranteed MPSoCs,

    W. Liu, J. Yi, M. Li, P. Chen, and L. Yang, “Energy-efficient application mapping and scheduling for lifetime guaranteed MPSoCs,”IEEE Trans. Comput. Aided Des. Integr. Circuits Syst., vol. 38, no. 1, pp. 1–14, Jan. 2019

    2019

  24. [24]

    Energy efficient, real-time and reliable task deployment on NoC-based multicores with DVFS,

    L. Mo, Q. Zhou, A. Kritikakou, and J. Liu, “Energy efficient, real-time and reliable task deployment on NoC-based multicores with DVFS,” in Proc. Des. Autom. Test Eur. Conf. Exhib. (DATE), 2022, pp. 1347–1352

    2022

  25. [25]

    Energy optimized task mapping for reliable and real-time networked systems,

    L. Mo, Q. Zhou, A. Kritikakou, and X. Cao, “Energy optimized task mapping for reliable and real-time networked systems,”ACM Trans. Sen. Netw., vol. 19, no. 4, pp. 1–26, Nov. 2023

    2023

  26. [26]

    An energy-aware multistages hybrid scheduling approach for IoT workflow applications with relia- bility constraint in cloud computing systems,

    L. Ye, L. Yang, X. Zhao, and Y . Xia, “An energy-aware multistages hybrid scheduling approach for IoT workflow applications with relia- bility constraint in cloud computing systems,”IEEE Internet Things J., vol. 12, no. 15, pp. 29 670–29 681, Aug. 2025

    2025

  27. [27]

    Energy-efficient task scheduling and resource allocation in edge-heterogeneous computing systems using multiobjective optimization,

    Q. Jiang, X. Xin, T. Zhang, and K. Chen, “Energy-efficient task scheduling and resource allocation in edge-heterogeneous computing systems using multiobjective optimization,”IEEE Internet Things J., vol. 12, no. 17, pp. 36 747–36 764, Sep. 2025

    2025

  28. [28]

    An energy-efficient, QoS-aware and cost-effective scheduling approach for real-time workflow appli- cations in cloud computing systems utilizing DVFS and approximate computations,

    G. L. Stavrinides and H. D. Karatza, “An energy-efficient, QoS-aware and cost-effective scheduling approach for real-time workflow appli- cations in cloud computing systems utilizing DVFS and approximate computations,”Future Gener. Comput. Syst., vol. 96, pp. 216–226, Jul. 2019

    2019

  29. [29]

    ASDYS: Dynamic scheduling using active strategies for multifunctional mixed-criticality cyber-physical systems,

    Y . Bai, Y . Huang, G. Xie, R. Li, and W. Chang, “ASDYS: Dynamic scheduling using active strategies for multifunctional mixed-criticality cyber-physical systems,”IEEE Trans. Ind. Inf., vol. 17, no. 8, pp. 5175– 5184, Aug. 2021

    2021

  30. [30]

    Multi-objective scheduling of extreme data scientific workflows in fog,

    V . De Maio and D. Kimovski, “Multi-objective scheduling of extreme data scientific workflows in fog,”Future Gener. Comput. Syst., vol. 106, pp. 171–184, May 2020

    2020

  31. [31]

    MCDS: AI augmented workflow scheduling in mobile edge cloud computing systems,

    S. Tuli, G. Casale, and N. R. Jennings, “MCDS: AI augmented workflow scheduling in mobile edge cloud computing systems,”IEEE Trans. Parallel Distrib. Syst., vol. 33, no. 11, pp. 2794–2807, Nov. 2022

    2022

  32. [32]

    A multi-objective approach for workflow scheduling in heterogeneous en- vironments,

    H. M. Fard, R. Prodan, J. J. D. Barrionuevo, and T. Fahringer, “A multi-objective approach for workflow scheduling in heterogeneous en- vironments,” inProc. IEEE/ACM Int. Symp. Clust. Cloud Grid Comput. (CCGrid), 2012, pp. 300–309

    2012

  33. [33]

    Dynamic DAG scheduling on multiprocessor systems: Reliability, energy, and makespan,

    J. Huang, R. Li, X. Jiao, Y . Jiang, and W. Chang, “Dynamic DAG scheduling on multiprocessor systems: Reliability, energy, and makespan,”IEEE Trans. Comput. Aided Des. Integr. Circuits Syst., vol. 39, no. 11, pp. 3336–3347, Nov. 2020

    2020

  34. [34]

    Improved many-objective particle swarm optimization algorithm for scientific workflow scheduling in cloud computing,

    S. Saeedi, R. Khorsand, S. Ghandi Bidgoli, and M. Ramezanpour, “Improved many-objective particle swarm optimization algorithm for scientific workflow scheduling in cloud computing,”Comput. Ind. Eng., vol. 147, p. 106649, Sep. 2020

    2020

  35. [35]

    Reliability, rental-cost and energy- aware multi-workflow scheduling on multi-cloud systems,

    A. Taghinezhad-Niar and J. Taheri, “Reliability, rental-cost and energy- aware multi-workflow scheduling on multi-cloud systems,”IEEE Trans. Cloud Comput., vol. 11, no. 3, pp. 2681–2692, Jul. 2023

    2023

  36. [36]

    Cost-aware reliability task scheduling of automotive cyber-physical systems,

    Y . Tang, Y . Yuan, and Y . Liu, “Cost-aware reliability task scheduling of automotive cyber-physical systems,”Microprocess. Microsyst., vol. 87, p. 103507, Nov. 2021

    2021

  37. [37]

    Minimizing redundancy to satisfy reliability requirement for a parallel application 18 on heterogeneous service-oriented systems,

    G. Xie, G. Zeng, Y . Chen, Y . Bai, Z. Zhou, R. Li, and K. Li, “Minimizing redundancy to satisfy reliability requirement for a parallel application 18 on heterogeneous service-oriented systems,”IEEE Trans. Serv. Comput., vol. 13, no. 5, pp. 871–886, Sep. 2020

    2020

  38. [38]

    Latency and reliability-aware workload assignment in IoT networks with mobile edge clouds,

    N. Kherraf, S. Sharafeddine, C. M. Assi, and A. Ghrayeb, “Latency and reliability-aware workload assignment in IoT networks with mobile edge clouds,”IEEE Trans. Netw. Serv. Manag., vol. 16, no. 4, pp. 1435–1449, Dec. 2019

    2019

  39. [39]

    A bioinspired method for optimal task scheduling in fog-cloud environment,

    F. Anka, G. G. Tejani, S. K. Sharma, and M. Baljon, “A bioinspired method for optimal task scheduling in fog-cloud environment,”Comput. Model. Eng. Sci., vol. 142, no. 3, pp. 2691–2724, Mar. 2025

    2025

  40. [40]

    Gurobi Optimization, https://www.gurobi.com, acc. Jul. 8, 2025

    2025

  41. [41]

    Fault-tolerant mapping of real-time parallel applications under multiple DVFS schemes,

    M. Cui, A. Kritikakou, L. Mo, and E. Casseau, “Fault-tolerant mapping of real-time parallel applications under multiple DVFS schemes,” in Proc. IEEE Real-Time Embed. Technol. Appl. Symp. (RTAS), 2021, pp. 387–399

    2021

  42. [42]

    A close examination of performance and power characteristics of 4G LTE networks,

    J. Huang, F. Qian, A. Gerber, Z. M. Mao, S. Sen, and O. Spatscheck, “A close examination of performance and power characteristics of 4G LTE networks,” inProc. Int. Conf. Mob. Syst. Appl. Serv. (MobiSys), 2012, pp. 225–238

    2012

  43. [43]

    High-resolution image transmission from UA V to ground station for search and rescue missions planning,

    V . Papi´c, P. ˇSoli´c, A. Milan, S. Gotovac, and M. Poli ´c, “High-resolution image transmission from UA V to ground station for search and rescue missions planning,”Appl. Sci., vol. 11, no. 5, pp. 1–19, Feb. 2021

    2021

  44. [44]

    Powertop, https://github.com/fenrus75/powertop, acc. Jul. 8, 2025

    2025

  45. [45]

    TGFF: Task graphs for free,

    R. P. Dick, D. L. Rhodes, and W. Wolf, “TGFF: Task graphs for free,” inProc. Int. Workshop Hardw./Softw. Codesign, 1998, pp. 97–101

    1998

  46. [46]

    Geekbench, https://www.geekbench.com, acc. Jul. 8, 2025

    2025

  47. [47]

    Open Benchmarking, https://openbenchmarking.org, acc. Jul. 8, 2025

    2025

  48. [48]

    Datasets of synthetic workflows for evaluating a multi-objective and multi-constrained scheduling approach for cyber-physical applications,

    A. Kouloumpris, G. L. Stavrinides, M. K. Michael, and T. Theocharides, “Datasets of synthetic workflows for evaluating a multi-objective and multi-constrained scheduling approach for cyber-physical applications,” Apr. 2024. [Online]. Available: https://doi.org/10.5281/zenodo.10978009 Andreas Kouloumprisgraduated top of his class receiving the BSc degree i...

    work page doi:10.5281/zenodo.10978009 2024