Towards a Distributed Federated Learning Aggregation Placement using Particle Swarm Intelligence

Amir Ali-Pour; julien gascon-samson; Laya Samizadeh; Sadra Bekrani

REVIEW 3 major objections 6 minor 25 references

Towards a Distributed Federated Learning Aggregation Placement using Particle Swarm Intelligence

T0 review · 3 major / 6 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 The paper claims that a particle swarm optimizer can choose which clients act as aggregators in hierarchical semi-decentralized federated learning using only the total processing delay as feedback, and that this placement is about 30…

desk verdict A reasonable black-box PSO placement idea for semi-decentralized FL, with a Docker proof-of-concept undermined by a single stochastic baseline draw and an unvalidated simulation model. read the letter →

arxiv 2504.16227 v1 pith:B5TBWE4F submitted 2025-04-22 cs.DC cs.LGcs.NEcs.NI

Amir Ali-Pour , Sadra Bekrani , Laya Samizadeh , Julien Gascon-Samson This is my paper · ORCID

classification cs.DCcs.LGcs.NEcs.NI

keywords federatedlearningsemi-decentralizedaggregationplacementparticleswarmoptimizationblack-boxtaskprocessingdelay

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 argues that aggregation placement in hierarchical semi-decentralized federated learning (SDFL) can be treated as a black-box optimization problem: observe only the total processing delay of each round, and let a particle swarm optimizer propose which clients should act as aggregators. The authors claim this removes the need for clients to report CPU, memory, bandwidth, or other internal telemetry, preserving privacy and reducing communication overhead. If correct, a lightweight optimizer can match or beat simple deterministic placement heuristics in real deployments, saving tens of minutes over a training run. The evidence is a simulated hierarchy with up to hundreds of clients plus a containerized implementation on an MQTT-based FL framework.

What carries the argument

The load-bearing object is the PSO particle as a placement vector combined with the TPD fitness model. A particle is a vector of client IDs assigned to aggregator slots; position updates use modular arithmetic, $x^{t+1}_i = (x^t_i + v^{t+1}_i) \% \text{client\_count}$, with velocity clamped to $V_{\max} = \max(1, D \times \text{velocity\_factor})$. The fitness function guiding the search is the negative TPD from Eqs. 6–7, which assumes each aggregator's delay equals the model data size plus the data sizes of its children divided by its processing speed, and that the round delay is the sum over hierarchy levels of the maximum such cluster delay. This synthetic model lets the optimizer evaluate candidate placements without any client telemetry.

What would settle it

Rerun the 10-container scenario many times with random, round-robin, and PSO-based placements; if PSO does not consistently beat the baselines by roughly 20–30 minutes, or if the measured per-round delays do not correlate with the Eq. 6–7 TPD predictions for the same placements, the central claim fails. A stronger test is to instrument the containers, record actual aggregation times for the placements PSO selects, and confirm those placements are indeed faster when deployed.

Watch

Extended reading notes

Core claim

The paper's central claim is that particle swarm optimization can find near-optimal aggregator assignments using only the total processing delay (TPD) as feedback. Each particle is a permutation of client IDs assigned to aggregator slots; after each round the coordinator measures the TPD, feeds it as fitness $f = -T$, updates velocities with the standard PSO rule, and rearranges the hierarchy for the next round. The TPD is computed bottom-up as the sum over hierarchy levels of the maximum cluster delay, where each cluster delay is $d_a = (\text{mdata}_a + \sum_{c \in \text{children}(a)} \text{mdata}_c) / \text{pspeed}_a$. In simulation with depth 3–5 and width 4–5, the optimizer converges within tens of iterations; in a 10-container deployment it converged after round 10, and over 50 rounds the PSO placement produced a total processing time roughly 30 minutes shorter than random placement and 20 minutes shorter than round-robin placement.

Load-bearing premise

The simulation fitness model, which assumes each aggregator's delay is the sum of model sizes in its subtree divided by processing speed and that the round delay is the sum of per-level maxima, is taken as a faithful proxy for real SDFL processing delay, yet it is not validated against measured delays from the container experiment, and that experiment ran only a single configuration.

Editorial extensions

If this is right

Aggregator selection no longer requires clients to expose internal performance data, reducing communication load and strengthening privacy.
In hierarchies with hundreds of clients, PSO converges in tens of iterations, making the approach feasible for dynamic role reassignment between rounds.
PSO-based placement beats random and round-robin placement in total processing time in the tested containerized deployment.
Because the optimizer is agnostic to the source of delay, the same loop can be reused if the fitness metric changes, provided the metric can be measured after each round.

Reading between the lines

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

The paper leaves untested whether PSO remains effective when node speeds drift over time; re-running the optimizer periodically could track such drift, but this is an extension the authors do not demonstrate.
A natural testable extension is to compare PSO against Bayesian optimization or genetic algorithms on the same TPD-only objective to see if PSO's claimed convergence advantage holds across a wider range of hierarchies.
The synthetic TPD model likely ignores communication latency between the broker and clients; adding a measured communication-delay term as a second black-box signal could improve placement in wide-area deployments, but that is speculation beyond the paper's simulations.
If the Eq. 6–7 model mispredicts real aggregation delay, PSO may converge to placements that look optimal in simulation but are not fastest in practice; validating the model against measured delays is the key next experiment.

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

Colleague —

Here's the short version: the paper proposes Flag-Swap, a PSO-based black-box optimizer that assigns aggregator roles in hierarchical semi-decentralized FL using only the measured total processing delay. That's a genuine, if modest, new application — the related work covers PSO for cloud scheduling and edge offloading, but not this particular placement problem. The authors integrate it into their own SDFLMQ framework and show in Docker that PSO converges after about ten rounds and then runs faster per round than their two built-in strategies.

What's good: the idea is clean and the implementation is real engineering. They avoid client telemetry entirely, which is a nice property for privacy and communication overhead. The simulation study spans different tree depths and widths and suggests the optimizer scales to hundreds of clients. The plots show typical PSO convergence behavior, and the writing is straightforward.

The soft spots are mostly in the evaluation. The headline claim — 43% faster than random, 32% faster than uniform — rests on a single 50-round Docker run with one PSO trajectory, one random draw, and one round-robin run. Random placement is stochastic; a single draw can be arbitrarily bad. There are no seeds, no replicates, no error bars, so those percentages are anecdotes, not measurements. The stress-test note is right: that gap is load-bearing. If repeated trials put PSO within noise of random, the central superiority claim collapses.

The simulation can't rescue it. The synthetic TPD model in Eq. 6-7 sums per-level maximum cluster delays as (model size + children sizes) / processing speed, and it's never validated against the measured Docker delays. It also omits communication and broker costs, which are likely significant in an MQTT-based system. Calling the converged value the "global minimum" without exhaustive search is an overstatement, though harmless if read as a heuristic claim. There's also no comparison with other black-box optimizers like a simple genetic algorithm or Bayesian optimization, which weakens the "PSO is a good choice" argument.

Minor things: the abstract's "43% minutes faster" is awkward, and the reliance on their own SDFLMQ framework is fine as a platform choice, but it means the result is at least partly tied to that specific system.

Who's this for? Researchers working on orchestration for hierarchical FL, or anyone curious about black-box placement in distributed systems. It's a proof-of-concept, not a settled result, but the problem is real and the approach is worth knowing about.

My call: send it to peer review, but require major revision — repeated trials with seeds and confidence intervals, validation of the simulation model against real measurements, and a softened claim about global optimality. The core idea stands; the evidence needs to catch up.

Referee Report

3 major / 6 minor

Summary. The manuscript proposes Flag-Swap, a PSO-based black-box optimizer for selecting aggregator placements in hierarchical semi-decentralized federated learning (SDFL). The method uses only the total processing delay per FL round as the fitness signal, avoiding client-side telemetry. The authors present simulations over synthetic tree hierarchies and a Docker-based deployment on the SDFLMQ framework, comparing PSO with random and round-robin placement. The paper claims that PSO converges quickly and reduces total processing time by about 43% over random placement and 32% over uniform placement.

Significance. If the performance claims hold, the paper would demonstrate that a black-box optimizer can improve aggregation placement without requiring client telemetry, which is a useful and potentially privacy-preserving direction for SDFL systems. The paper includes both simulation and a real Docker deployment, and the underlying SDFLMQ framework is publicly available, which supports reproducibility of the platform if not the additional optimizer code. The contribution is conceptually plausible, but the current experimental evidence is insufficient to support the headline quantitative claims; the reported gains rest on a single run per strategy and an unvalidated synthetic delay model.

major comments (3)

[Section IV.C, Fig. 4] The central quantitative claim (about 43% and 32% faster total processing time) rests on a single Docker experiment with one run per placement strategy. Random placement is stochastic, and a single random draw is not a reliable baseline; the paper reports no seeds, no repeated trials, and no confidence intervals, so the reported improvements are not statistically established. I request several independent runs of each strategy, including multiple random seeds for the random baseline, with reported distributions or paired statistical tests.
[Section IV.B and Eqs. (6)-(7)] The simulation TPD model sums per-level maximum cluster delays computed as (model size plus children's sizes) divided by processing speed, but this synthetic model is never validated against the measured delays from the Docker experiment, and it omits communication and broker costs. Consequently, the simulation's claim that PSO reaches 'the global minimum TPD' is unsupported: for this discrete permutation search space, convergence of the swarm does not guarantee a global optimum without exhaustive enumeration or a lower-bound argument. Please either validate the model against real measurements, soften the optimality claim, or provide exhaustive search results for small instances.
[Section IV.C] The reported total processing time includes PSO's exploration overhead in the first approximately 10 rounds. This is a legitimate system-level metric, but the paper does not separate warm-up/exploration cost from steady-state per-round advantage, so the headline improvement conflates convergence cost with placement quality. Please report both total-time and steady-state per-round metrics, and analyze sensitivity to the number of rounds.

minor comments (6)

[Abstract and Section IV.C] The quantitative statements are inconsistent: the abstract says 'about 43% minutes faster than random placement, and 32% minutes faster than uniform placement,' while Section IV.C says 'around 30 minutes faster than random-based placement, and around 20 minutes faster than uniform-based placement.' Please reconcile these numbers.
[Section IV.A] The term 'Agtrainers' appears to be a typo; it should likely read 'Aggregators.' Also, in the introduction, 'specially' should be 'spatially' if that is the intended meaning.
[Section IV.B, Fig. 3] The figure captions say 'Client number 81' etc., but the text describes the number of clients. Please clarify that the caption refers to the total client count.
[Algorithm 1 and Section III.C.3] The duplicate-resolution step ('Duplicates are resolved by incrementing until a unique client ID is found') is unclear and may bias the search; please specify the mechanism and its effect on the velocity/position encoding.
[References] Reference [16] is incomplete: it lacks authors and title. Please provide the full citation for the SDFLMQ Python source code.
[Section IV.C] The statement that 'PSO-based placement was able to converge after the 10th round' should be accompanied by a precise convergence criterion, as Fig. 4 shows per-round processing time varying after that point.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the optimizer minimizes the same TPD metric used for evaluation in simulation, and the Docker-based comparison uses measured processing times independently of that fitness model.

full rationale

The paper's derivation chain is empirical rather than formal. In simulation, the fitness function is the Total Processing Delay defined by Eqs. 6-7, and PSO iteratively minimizes that same function; evaluating an optimizer against its own objective is standard black-box optimization, not circular reasoning. The main empirical claim, about Docker-based performance, is supported by measured per-round processing times, not by substituting values into the simulation model, so the Docker claim does not reduce to the fitted or synthetic fitness function. The self-citation to SDFLMQ [21] supplies the experimental framework, but the comparison with random and round-robin placement is an independent measurement within that framework; it is not a theorem whose conclusion is assumed in its premises. No fitted parameter is relabeled as a prediction, no uniqueness theorem is imported from prior work, and no ansatz is smuggled in through a citation. Concerns about a single Docker run, lack of repeated trials or seeds, and the unvalidated synthetic TPD model are statistical and external-validity issues, not circularity. Therefore the honest finding is no significant circularity.

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

The method introduces no invented entities and fits no parameters to data. The PSO hyperparameters and simulation settings are hand-chosen but not tuned to the results. The main conceptual debt is the assumption that the synthetic TPD model captures real SDFL delay structure, and that the single Docker scenario generalizes.

free parameters (2)

PSO hyperparameters = w=0.01, c1=0.01, c2=1, velocity factor=0.1, P={5,10}, iterations=100
Chosen by hand from common PSO practice; not fitted to the data, but they affect convergence speed and are part of the method specification.
Simulation model parameters = mdatasize=5, pspeed in [5,15], mem in [10,50], 2 trainers per leaf aggregator
Arbitrary testbed settings for the synthetic simulation; results are specific to these ranges and the balanced-tree topology.

assumptions (4)

domain assumption The synthetic TPD model (Eq. 6-7), summing per-level maximum cluster delay, faithfully represents the delay bottleneck of a real SDFL system.
Invoked in Section IV-A as the fitness function; without this, simulation convergence does not transfer to real-world performance.
domain assumption The SDFL hierarchy is a balanced tree with fixed depth D and width W, and clients have static attributes during optimization.
Used throughout Section IV; real systems may have dynamic topologies and changing resources.
domain assumption The PSO position update with modulo arithmetic and duplicate incrementing yields feasible, sufficiently diverse placements.
Eq. 4 and Algorithm 1; no analysis of repair bias or exploration quality in the discrete permutation space.
domain assumption The SDFLMQ framework (self-cited in [21]) supports dynamic role reassignment at each round without significant overhead.
The real-world experiment in Section IV-C depends on this capability; the framework is the authors' own and not independently evaluated here.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Towards a Distributed Federated Learning Aggregation Placement using Particle Swarm Intelligence." pith.science (2026). https://pith.science/paper/B5TBWE4F

@misc{pith2026250416227,
  author       = {Pith},
  title        = {Pith review of: Towards a Distributed Federated Learning Aggregation Placement using Particle Swarm Intelligence},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B5TBWE4F}},
  note         = {Machine review of arXiv:2504.16227}
}

read the original abstract

Federated learning has become a promising distributed learning concept with extra insurance on data privacy. Extensive studies on various models of Federated learning have been done since the coinage of its term. One of the important derivatives of federated learning is hierarchical semi-decentralized federated learning, which distributes the load of the aggregation task over multiple nodes and parallelizes the aggregation workload at the breadth of each level of the hierarchy. Various methods have also been proposed to perform inter-cluster and intra-cluster aggregation optimally. Most of the solutions, nonetheless, require monitoring the nodes' performance and resource consumption at each round, which necessitates frequently exchanging systematic data. To optimally perform distributed aggregation in SDFL with minimal reliance on systematic data, we propose Flag-Swap, a Particle Swarm Optimization (PSO) method that optimizes the aggregation placement according only to the processing delay. Our simulation results show that PSO-based placement can find the optimal placement relatively fast, even in scenarios with many clients as candidates for aggregation. Our real-world docker-based implementation of Flag-Swap over the recently emerged FL framework shows superior performance compared to black-box-based deterministic placement strategies, with about 43% minutes faster than random placement, and 32% minutes faster than uniform placement, in terms of total processing time.

Figures

Figures reproduced from arXiv: 2504.16227 by the authors.

**Figure 2.** Proposed PSO-based aggregation placement in SDFL. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Simulation results of PSO optimization in aggregation placement in SDHFL. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Comparing aggregation placement using Random, [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 22 canonical work pages

[1]

Nguyen, Dinh C, Ding, Ming, Pathirana, Pubudu N, Seneviratne, Aruna, Li, Jun, Poor, H Vincent, ”Federated learning for internet of things: A comprehensive survey,”IEEE Communications Surveys & Tutorials , vol. 23, no. 3, pp. 1622–1658, 2021

work page 2021
[2]

Zhang, Tuo, Gao, Lei, He, Chaoyang, Zhang, Mi, Krishnamachari, Bhaskar, Avestimehr, A Salman, ”Federated learning for the internet of things: Applications, challenges, and opportunities,” IEEE Internet of Things Magazine , vol. 5, no. 1, pp. 24–29, 2022

work page 2022
[3]

Lim, Wei Yang Bryan, Luong, Nguyen Cong, Hoang, Dinh Thai, Jiao, Yutao, Liang, Ying-Chang, Yang, Qiang, Niyato, Dusit, Miao, Chunyan, ”Federated learning in mobile edge networks: A comprehensive survey,” IEEE communications surveys & tutorials, vol. 22, no. 3, pp. 2031–2063, 2020

work page 2020
[4]

Ji, Xiuzhao, Tian, Jie, Zhang, Haixia, Wu, Dalei, Li, Tiantian, ”Joint device selection and bandwidth allocation for cost-efficient federated learning in industrial internet of things,” IEEE Internet of Things Journal, vol. 10, no. 10, pp. 9148–9160, 2023

work page 2023
[5]

Guo, Yinghao, Zhao, Zichao, He, Ke, Lai, Shiwei, Xia, Junjuan, Fan, Lisheng, ”Efficient and flexible management for industrial internet of things: A federated learning approach,” Computer Networks , vol. 192, pp. 108122, 2021

work page 2021
[6]

Bonawitz, Keith, ”Towards federated learning at scale: Syste m design,” arXiv preprint arXiv:1902.01046 , 2019

arXiv 1902
[7]

Beltr ´an, Enrique Tom ´as Mart´ınez, P ´erez, Mario Quiles, S ´anchez, Pedro Miguel S ´anchez, Bernal, Sergio L ´opez, Bovet, G ´erˆome, P ´erez, Manuel Gil, P´erez, Gregorio Mart´ınez, Celdr´an, Alberto Huertas, ”Decentralized federated learning: Fundamentals, state of the art, frameworks, trends, and challenges,” IEEE Communications Surveys & Tutorials , 2023

work page 2023
[8]

Luo, Siqi, Chen, Xu, Wu, Qiong, Zhou, Zhi, Yu, Shuai, ”HFEL: Joint edge association and resource allocation for cost-efficient hierarchical federated edge learning,” IEEE Transactions on Wireless Communica- tions, vol. 19, no. 10, pp. 6535–6548, 2020

work page 2020

Show all 25 references

[9]

Xu, Bo, Zhao, Haitao, Cao, Haotong, Garg, Sahil, Kaddoum, Georges, Hassan, Mohammad Mehedi, ”Edge aggregation placement for semi- decentralized federated learning in Industrial Internet of Things,” Future Generation Computer Systems , vol. 150, pp. 160–170, 2024. This preprint...

2024
[10]

111–139, 2021

Ziller, Alexander, Trask, Andrew, Lopardo, Antonio, Szymkow, Ben- jamin, Wagner, Bobby, Bluemke, Emma, Nounahon, Jean-Mickael, Passerat-Palmbach, Jonathan, Prakash, Kritika, Rose, Nick, others, ”Pysyft: A library for easy federated learning,” Federated Learning Systems: Toward...

2021
[11]

He, Chaoyang, Li, Songze, So, Jinhyun, Zeng, Xiao, Zhang, Mi, Wang, Hongyi, Wang, Xiaoyang, Vepakomma, Praneeth, Singh, Abhishek, Qiu, Hang, others, ”Fedml: A research library and benchmark for federated machine learning,” arXiv preprint arXiv:2007.13518 , 2020

2007 arXiv
[12]

Beutel, Daniel J, Topal, Taner, Mathur, Akhil, Qiu, Xinchi, Fernandez- Marques, Javier, Gao, Yan, Sani, Lorenzo, Li, Kwing Hei, Par- collet, Titouan, de Gusm ˜ao, Pedro Porto Buarque, others, ”Flower: A friendly federated learning research framework,” arXiv preprint arXiv:2007...

2007 arXiv
[13]

Lin, Frank Po-Chen, Hosseinalipour, Seyyedali, Azam, Sheikh Shams, Brinton, Christopher G, Michelusi, Nicolo, ”Semi-decentralized feder- ated learning with cooperative D2D local model aggregations,” IEEE Journal on Selected Areas in Communications , vol. 39, no. 12, pp. 3851– ...

2021
[14]

Beltr ´an, Enrique Tom ´as Mart ´ınez, G ´omez, ´Angel Luis Perales, Feng, Chao, S ´anchez, Pedro Miguel S ´anchez, Bernal, Sergio L ´opez, Bovet, G´erˆome, P´erez, Manuel Gil, P ´erez, Gregorio Mart´ınez, Celdr´an, Alberto Huertas, ”Fedstellar: A platform for decentralized fe...

2024
[15]

Sun, Yuchang, Shao, Jiawei, Mao, Yuyi, Wang, Jessie Hui, Zhang, Jun, ”Semi-decentralized federated edge learning with data and device het- erogeneity,” IEEE Transactions on Network and Service Management , vol. 20, no. 2, pp. 1487–1501, 2023

2023
[16]

, ”SDFLMQ Python Source code,” , 2025

2025
[17]

You, Qian, Tang, Bing, ”Efficient task offloading using particle swarm optimization algorithm in edge computing for industrial internet of things,” Journal of Cloud Computing , vol. 10, pp. 1–11, 2021

2021
[18]

Fu, Xueliang, Sun, Yang, Wang, Haifang, Li, Honghui, ”Task scheduling of cloud computing based on hybrid particle swarm algorithm and genetic algorithm,” Cluster Computing , vol. 26, no. 5, pp. 2479–2488, 2023

2023
[19]

Ben Alla, Hicham, Ben Alla, Said, Touhafi, Abdellah, Ezzati, Abdellah, ”A novel task scheduling approach based on dynamic queues and hybrid meta-heuristic algorithms for cloud computing environment,” Cluster Computing, vol. 21, no. 4, pp. 1797–1820, 2018

2018
[20]

Nabi, Said, Ahmad, Masroor, Ibrahim, Muhammad, Hamam, Habib, ”AdPSO: adaptive PSO-based task scheduling approach for cloud com- puting,” Sensors, vol. 22, no. 3, pp. 920, 2022

2022
[21]

Amir Ali-Pour, Julien Gascon-Samson, ”SDFLMQ: A Semi- Decentralized Federated Learning Framework over MQTT,” arXiv preprint, 2025

2025
[22]

Meunier, Laurent, Rakotoarison, Herilalaina, Wong, Pak Kan, Roziere, Baptiste, Rapin, J ´er´emy, Teytaud, Olivier, Moreau, Antoine, Doerr, Carola, ”Black-box optimization revisited: Improving algorithm selec- tion wizards through massive benchmarking,” IEEE Transactions on Evo...

2021
[23]

Boveiri, Hamid Reza, Khayami, Raouf, ”On the performance of meta- heuristics: A different perspective,” arXiv preprint arXiv:2001.08928 , 2020

2001 arXiv
[24]

1–6, 2020

Liu, Lumin, Zhang, Jun, Song, SH, Letaief, Khaled B, ”Client-edge- cloud hierarchical federated learning,” , pp. 1–6, 2020

2020
[25]

3–15, 2009

Auger, Anne, Hansen, Nikolaus, Perez Zerpa, Jorge M, Ros, Raymond, Schoenauer, Marc, ”Experimental Comparisons of Derivative Free Op- timization Algorithms: (Invited Talk),” , pp. 3–15, 2009. This preprint version made available under the arXiv.org perpetual, non-exclusive license

2009

Pith tools

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

[1] [1]

Nguyen, Dinh C, Ding, Ming, Pathirana, Pubudu N, Seneviratne, Aruna, Li, Jun, Poor, H Vincent, ”Federated learning for internet of things: A comprehensive survey,”IEEE Communications Surveys & Tutorials , vol. 23, no. 3, pp. 1622–1658, 2021

work page 2021

[2] [2]

Zhang, Tuo, Gao, Lei, He, Chaoyang, Zhang, Mi, Krishnamachari, Bhaskar, Avestimehr, A Salman, ”Federated learning for the internet of things: Applications, challenges, and opportunities,” IEEE Internet of Things Magazine , vol. 5, no. 1, pp. 24–29, 2022

work page 2022

[3] [3]

Lim, Wei Yang Bryan, Luong, Nguyen Cong, Hoang, Dinh Thai, Jiao, Yutao, Liang, Ying-Chang, Yang, Qiang, Niyato, Dusit, Miao, Chunyan, ”Federated learning in mobile edge networks: A comprehensive survey,” IEEE communications surveys & tutorials, vol. 22, no. 3, pp. 2031–2063, 2020

work page 2020

[4] [4]

Ji, Xiuzhao, Tian, Jie, Zhang, Haixia, Wu, Dalei, Li, Tiantian, ”Joint device selection and bandwidth allocation for cost-efficient federated learning in industrial internet of things,” IEEE Internet of Things Journal, vol. 10, no. 10, pp. 9148–9160, 2023

work page 2023

[5] [5]

Guo, Yinghao, Zhao, Zichao, He, Ke, Lai, Shiwei, Xia, Junjuan, Fan, Lisheng, ”Efficient and flexible management for industrial internet of things: A federated learning approach,” Computer Networks , vol. 192, pp. 108122, 2021

work page 2021

[6] [6]

Bonawitz, Keith, ”Towards federated learning at scale: Syste m design,” arXiv preprint arXiv:1902.01046 , 2019

arXiv 1902

[7] [7]

Beltr ´an, Enrique Tom ´as Mart´ınez, P ´erez, Mario Quiles, S ´anchez, Pedro Miguel S ´anchez, Bernal, Sergio L ´opez, Bovet, G ´erˆome, P ´erez, Manuel Gil, P´erez, Gregorio Mart´ınez, Celdr´an, Alberto Huertas, ”Decentralized federated learning: Fundamentals, state of the art, frameworks, trends, and challenges,” IEEE Communications Surveys & Tutorials , 2023

work page 2023

[8] [8]

Luo, Siqi, Chen, Xu, Wu, Qiong, Zhou, Zhi, Yu, Shuai, ”HFEL: Joint edge association and resource allocation for cost-efficient hierarchical federated edge learning,” IEEE Transactions on Wireless Communica- tions, vol. 19, no. 10, pp. 6535–6548, 2020

work page 2020