Pith. sign in

REVIEW 5 major objections 5 minor 36 references

Raft Distributed System for Multi-access Edge Computing Sharing Resources

T0 review · 5 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read A DDPG-trained policy can choose Raft's leader to cut latency in edge-computing clusters.

desk verdict An incomplete early draft that asserts a DDPG-Raft combination for MEC but never defines how the two connect, and the results section has no data; not ready for any serious referee. read the letter →

arxiv 2412.16774 v1 pith:EVTZZMDE submitted 2024-12-21 cs.DC cs.ET

Zain Khaliq , Ahmed Refaey Hussein This is my paper · ORCID
classification cs.DCcs.ET
keywords RaftconsensusMulti-accessEdgeComputingDDPGleaderelectionHyperledgerFabriclatencyoptimizationreinforcementlearningblockchain
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 tries to establish that a Multi-access Edge Computing (MEC) cluster sharing tasks through a Raft-based private blockchain can reduce overall latency by replacing Raft's randomized leader election with a Deep Deterministic Policy Gradient (DDPG) agent that picks the most resourceful node as leader. It models the cluster as a Hyperledger Fabric network where shortlisted edge nodes run an election, and it derives equations for the three main latency sources: task migration from the cloud, block generation, and block-consensus communication. The DDPG actor-critic is trained to choose an action (effectively the leader choice) by maximizing reward over 500 episodes, with explicit action refinement through the critic's Q-function. The paper concludes that the trained policy yields low standard deviation and high reward values, which it interprets as an optimal policy gradient that improves the system's responsiveness.

What carries the argument

The central object is Deep Deterministic Policy Gradient (DDPG), an actor-critic reinforcement learning method. The actor maps a state $S$ to a continuous action, exploration noise $g$ is added to collect experience, and the critic $Q(S, A)$ provides a value signal; target networks and soft updates stabilize training. The paper augments the standard algorithm with an explicit action-refinement step, $A^* = \arg\max_{\tilde{A}\in\mathcal{A}} \hat{Q}(S, \tilde{A})$, and connects this to Raft's leader election so that the elected leader is, by training, the most resourceful node. This machinery replaces the randomized election timeout that normally drives Raft's leader election.

What would settle it

Run a multi-node cluster under the learned election policy in a partially synchronous network with injected message delays and a partitioned node; if the system ever records two leaders for the same term, or a term with no elected leader despite a majority of live nodes, the central claim is refuted. This can be checked by instrumenting Raft's RequestVotes and AppendEntries logs across many randomized network schedules.

Watch

Extended reading notes

Core claim

The central claim is that a DDPG-based resource-training algorithm, when inserted into Raft's election step, elects the most resourceful edge node in the cluster as the leader and thereby lowers end-to-end task latency compared with conventional Raft's randomized election timeout. The proposed mechanism is an actor-critic loop: the actor network outputs a candidate action from a Gaussian-perturbed policy, the critic refines that action via $A^* = \arg\max_{\tilde{A}\in\mathcal{A}} \hat{Q}(S, \tilde{A})$, and policy-gradient updates adjust the actor toward higher expected reward. Latency is quantified by three equations covering migration ($\tau^{\mathrm{TML}}_l$), block generation ($\tau^{\mathrm{CLBG}}_l$), and consensus over the wireless links ($\tau^{\mathrm{RCB}}_l$), and the DDPG objective is presented as optimizing those quantities. The paper reports training curves (average reward per trial) and moving averages as evidence that the learned policy is stable and improving.

Load-bearing premise

The load-bearing premise is that a machine-learned election policy still obeys Raft's rule of at most one leader per term and always eventually elects one, even when network delays are unpredictable.

Editorial extensions

If this is right

  • If correct, Raft-based MEC clusters can replace randomized election timeouts with a resource-aware policy, lowering the expected latency of each task offload.
  • The leader elected by the DDPG policy should, on average, be the node with the best combination of compute speed, bandwidth, and channel quality, so task execution and block generation happen on the most capable node.
  • Because the policy is trained across episodes, the cluster could adapt to changing load without re-architecting the consensus protocol.
  • The explicit latency equations (migration, block generation, consensus) give a measurable target: a deployed system could monitor those quantities to verify the claimed improvement.

Reading between the lines

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

  • The paper does not demonstrate that a learned election policy preserves Raft's safety and liveness; a direct test would be to run the learned policy under adversarial network delays and verify that no term ever has two leaders and that an election always terminates.
  • The auction-theory framing is left implicit; one could formalize leader election as a bid-based auction where each node's 'bid' is its predicted latency contribution, making the resource-optimality claim precise and testable.
  • The reported results are DDPG training curves (reward per episode), not end-to-end latency comparisons; a natural extension is to measure actual task latency with and without DDPG on the same cluster under identical traffic.
  • DDPG is normally designed for continuous control, while leader election is a discrete choice; implementing the paper's proposal in practice would require either a discrete-action variant or an explicit continuous relaxation, which the paper does not discuss.
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.

Referee Report

5 major / 5 minor

Summary. The paper proposes combining the Raft consensus algorithm with a Hyperledger Fabric-based multi-access edge computing (MEC) system for task sharing, and claims that a Deep Deterministic Policy Gradient (DDPG) agent can optimize the system by electing the most resourceful edge node as the Raft leader, thereby reducing overall latency. The manuscript reviews Raft background, introduces a latency decomposition in Section IV (migration latency, block-generation latency, and block-consensus latency), and then presents a DDPG training loop in Section V. The central claim, however, is never made precise: the state, action, reward, and transition function of the DDPG agent are not connected to Raft's election mechanism, and the results section refers to figures and empirical comparisons that are not present in the manuscript. The paper therefore does not currently provide a well-defined DDPG-based leader-election algorithm, nor evidence that such an algorithm would reduce latency or preserve Raft's safety and liveness.

Significance. If the paper had supplied a concrete mapping from DDPG state/action/reward to Raft leader election, together with a correctness argument that the learned policy preserves Raft's election safety and liveness, and an experimental comparison showing lower latency than randomized Raft election, the result would be a useful contribution to MEC resource sharing and blockchain-based edge systems. The manuscript does provide a readable summary of Raft and a plausible latency decomposition in Eqs. (1)-(5), which are useful background. However, the proposed DDPG contribution is not defined as an algorithm, and the reported results consist of absent figures and unquantified claims. As submitted, the central claim is unsupported, so the significance of the work cannot be assessed on the basis of the manuscript.

major comments (5)
  1. [Section V-A and Algorithm 1] The paper never defines the mapping from DDPG to Raft leader election. The initial state S0 = {0,...,0} is not specified in terms of node resources, Raft state, or the latency equations; the action A is sampled from a Gaussian distribution with no stated effect on election timeout, term number, or vote requests; the reward R is never defined; and the transition function f(S,A) in Eq. (8) is left as an unspecified function. Consequently, the claim in Section I that DDPG 'allows it to elect the most resourceful node in the cluster as its leader' is not a well-formed algorithmic claim. The authors need to define the state features, action semantics, reward, and transition in terms of the Raft protocol and Eqs. (1)-(5).
  2. [Section III-A and Section I] Replacing Raft's randomized election timeout with a learned policy must preserve Raft's Election Safety (at most one leader per term) and Liveness (eventual leader election), as established in Ongaro and Ousterhout [24]. The manuscript provides no invariant, no proof, and no simulation showing that a DDPG-learned policy under partial synchrony satisfies these properties. Without such an argument, the proposed system's correctness is unverified, and the claim that the system improves reliability and avoids task failure is unsupported.
  3. [Section V-D] The Results subsection claims that Fig. 1 and Fig. 2 show average reward per trial and reward per episode, and that the training and testing yielded low standard deviations and maximized reward values. However, no figures, numerical latency values, baseline comparisons with standard Raft, confidence intervals, or error analyses are present in the manuscript. The conclusion in Section VI that 'through the DDPG model, we have optimized our proposed system' is therefore not supported by any empirical evidence.
  4. [Equations (11)-(13) and Table I] The parameter roles of θ and ψ are inconsistent. Table I identifies θ as parameters from the actor network and ψ as parameters from the critic network, but Eq. (11) updates θ with the critic loss ∇L, Eq. (12) computes the actor gradient as ∂Q/∂ψ, and Eq. (6) writes A = π(S | ψ). The text also mentions learning rates α1 and α2 without clearly assigning them to actor or critic. These inconsistencies make the training procedure irreproducible as written.
  5. [Section V and Section IV] The DDPG optimization objective is never tied to the latency model. The reward R is not derived from Eqs. (1)-(5), and the optimization problem in Section V is not stated as minimizing τ_TML, τ_CLBG, or τ_RCB. As a result, even if the DDPG loop were well defined, there is no reason to believe that maximizing the implicit reward corresponds to reducing overall system latency, which is the paper's central claim.
minor comments (5)
  1. [Section IV] The manuscript contains placeholder text such as '(Fig coming soon)' and '(ADDD EQ 1 HERE)', which indicates that the paper is incomplete and needs to be finished before review.
  2. [Table I and Eq. (1)] The symbol table defines τ_TML with |Zl(τ)|, but Eq. (1) uses |Zl(Γ)|; the notation should be made consistent.
  3. [Section V-D] The text contains the typo 'DDGP' and an extended analogy about commuting from work to home that is out of place in a technical results section.
  4. [Section V-A and Section V-D] The number of epochs is inconsistent: Section V-D states 100 epochs and 500 episodes, while Section V-A mentions '500 epochs in this paper' and Algorithm 1 uses Γ_max as the maximum iterations.
  5. [Section IV] The phrase 'term leader N from the n-th cluster' is unclear, and the auction-theory and game-theory concepts mentioned in the abstract and index terms are never developed in the body of the paper.

Circularity Check

2 steps flagged · score 6.0 of 10

The claimed DDPG latency optimization reduces to DDPG's own reward objective: the trained reward curves are presented as proof that the MEC/Raft system is optimized, making the 'optimal policy' true by definition.

  1. fitted input called prediction [Section V-D (Results and Discussion); Section VI (Conclusion)]
    "These results show that through the DDPG model, we have optimized our proposed system."

    The 'results' preceding this conclusion are reward-value statistics of the DDPG agent itself: Section V-D says 'we added the distributions of the different reward values to find the model that resulted in the best reward value with the greatest number of times.' In Algorithm 1, reward R is the direct training objective (target y = R + lambda*Q', actor update ∇J = ∂Q/∂A · ∂π/∂ψ). A policy that maximizes reward is therefore optimized by construction. Presenting those same reward curves as evidence that the Raft/MEC system is optimized reports the fitted training target as an independent system-level prediction, with no comparison to the latency equations (1)-(5) or to a baseline Raft system.

  2. self definitional [Section V-A (Training), Eq. (7); Section VI (Conclusion)]
    "We declare our final output as π∗, which is the optimal policy. The optimal policy represents the best possible action for the actor to perform in our DDPG model."

    Section V-A defines π* as the DDPG model's own optimal policy, and Eq. (7) obtains it as A* = arg max_{Ã∈A} Q(S,Ã). The paper then concludes 'These results show that through the DDPG model, we have optimized our proposed system' and describes 'an optimal policy gradient.' No equation maps the DDPG action A to a Raft election timeout, term, or vote, and no equation maps the reward R to the MEC latency expressions (1)-(5). The claimed optimality is therefore equivalent to the algorithm's own definition; the Raft/MEC system-level improvement is stipulated rather than derived.

full rationale

The only quantitative support offered for the paper's central claim is in Section V-D: figures of average reward per trial and reward per episode for the DDPG agent. These are exactly the quantities the DDPG optimizer maximizes (Eqs. 6-13), so the conclusion that 'we have optimized our proposed system' is the training objective reported as a result—a fitted input called a prediction. I score this 6 because the paper's own claimed contribution (DDPG-based Raft leader election reducing latency) rests on that circular evidence. I did not score as circular the absence of any defined state/action/reward mapping from DDPG to Raft's election mechanism, nor the missing safety/liveness proof for a learned election policy under partial synchrony: those are correctness and support gaps, not reductions of outputs to inputs. The self-citations to prior work by the authors appear in the reference list, but none is load-bearing for the DDPG-to-Raft claim, so they do not increase the score; the Raft consensus background is an external standard result [24]. No machine-checked, code-reproduced, or externally benchmarked derivation is offered for the central latency-improvement claim.

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

The central proposal rests on several free choices (reward, state, action, hyperparameters) that are never specified, and on an unexamined assumption that a learned policy can replace Raft's randomized election without breaking consensus guarantees. No new entities are postulated beyond the generic system components.

free parameters (5)
  • DDPG learning rates alpha1 and alpha2
    Learning rates for critic and actor updates are mentioned in equations (11) and the actor update, but no values are given and there is no sensitivity analysis.
  • Discount factor lambda and soft update coefficient kappa
    Used in target value and target network updates; no values provided and no justification.
  • Number of epochs and episodes = 500 epochs (Section V-A) vs 100 epochs and 500 episodes (Section V-D)
    The paper contradicts itself on the number of epochs, and the chosen values are asserted without experimental justification.
  • Reward function R
    The DDPG training loop uses reward R, but the paper never specifies how reward relates to latency, task completion, or resource usage. The entire optimization is undefined.
  • State representation S and action space A
    The state is said to start as a zero vector and actions are sampled from a Gaussian distribution, but the physical meaning of states and actions in the MEC auction is never defined.
assumptions (5)
  • domain assumption Existence of a cluster of N edge nodes, each with a series of tasks.
    Stated in Section IV as the base network model.
  • domain assumption The edge nodes run on a Hyperledger Fabric private blockchain network.
    Assumed in Section IV with no justification for why Hyperledger Fabric is the right ledger.
  • domain assumption The latency equations (1)-(5) correctly model migration, block generation, and consensus latency.
    The equations are stated without derivation or citation, and are not connected to the DDPG reward.
  • ad hoc to paper DDPG converges to an optimal policy that generalizes to the MEC environment.
    The paper relies on DDPG convergence without proving any performance bound for this specific system; the training loop is generic.
  • ad hoc to paper Replacing Raft's randomized election with a DDPG-learned policy preserves Raft's safety and liveness.
    This is a load-bearing assumption that the paper never states or analyzes. Raft's election safety depends on randomized timeouts and term-based voting rules; learned policies could violate these guarantees.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Raft Distributed System for Multi-access Edge Computing Sharing Resources." pith.science (2026). https://pith.science/paper/EVTZZMDE

@misc{pith2026241216774,
  author       = {Pith},
  title        = {Pith review of: Raft Distributed System for Multi-access Edge Computing Sharing Resources},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EVTZZMDE}},
  note         = {Machine review of arXiv:2412.16774}
}
read the original abstract

Researchers all over the world are employing a variety of analysis approaches in attempt to provide a safer and faster solution for sharing resources via a Multi-access Edge Computing system. Multi-access Edge Computing (MEC) is a job-sharing method within the edge server network whose main aim is to maximize the pace of the computing process, resulting in a more powerful and enhanced user experience. Although there are many other options when it comes to determining the fastest method for computing processes, our paper introduces a rather more extensive change to the system model to assure no data loss and/or task failure due to any scrutiny in the edge node cluster. RAFT, a powerful consensus algorithm, can be used to introduce an auction theory approach in our system, which enables the edge device to make the best decision possible regarding how to respond to a request from the client. Through the use of the RAFT consensus, blockchain may be used to improve the safety, security, and efficiency of applications by deploying it on trustful edge base stations. In addition to discussing the best-distributed system approach for our (MEC) system, a Deep Deterministic Policy Gradient (DDPG) algorithm is also presented in order to reduce overall system latency. Assumed in our proposal is the existence of a cluster of N Edge nodes, each containing a series of tasks that require execution. A DDPG algorithm is implemented in this cluster so that an auction can be held within the cluster of edge nodes to decide which edge node is best suited for performing the task provided by the client.

Figures

Figures reproduced from arXiv: 2412.16774 by the authors.

Figure 1
Figure 1. Probability Distribution Function per Trial. [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 2
Figure 2. Reward Value per Episode. VI. CONCLUSION In conclusion, we proposed a consensus system that uses the Raft algorithm paired with private blockchains for sharing resources in a MEC system. The network architecture is a cluster of edge nodes via a Hyperledger Fabric network. This ledger uses a private blockchain as its base. To test the effectiveness of our proposed design, we observed its learning rate and policy grad… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 30 canonical work pages

  1. [16]

    A blockchain policy and charging control framework for roaming in cellular networks,

    A. Refaey, K. Hammad, S. Magierowski, and E. Hossain, “A blockchain policy and charging control framework for roaming in cellular networks,” IEEE Network, vol. 34, no. 3, pp. 170–177, 2020

    2020

  2. [24]

    In search of an understandable consensus algorithm,

    D. Ongaro and J. Ousterhout, “In search of an understandable consensus algorithm,” in 2014 USENIX Annual Technical Conference (USENIX ATC 14) . Philadelphia, PA: USENIX Association, Jun. 2014, pp. 305–319. [Online]. Available: https://www.usenix.org/conference/atc14/ technical-sessions/presentation/ongaro

    work page 2014

  3. [1]

    An adaptive framework for distributed complex applications development,

    I. Benyahia and M. Hilali, “An adaptive framework for distributed complex applications development,” in Proceedings. 34th International Conference on Technology of Object-Oriented Languages and Systems - TOOLS 34 , 2000, pp. 339–349

    work page 2000

  4. [2]

    On iot edge devices: Manifold unsupervised learning for som platforms,

    S. Asad and A. Refaey, “On iot edge devices: Manifold unsupervised learning for som platforms,” in 2021 IEEE International Conference on Imaging Systems and Techniques (IST) , 2021, pp. 1–5

    work page 2021

  5. [3]

    Distributed algorithm for lifetime maximization in a delay-tolerant wireless sensor network with a mobile sink,

    Y . Yun, Y . Xia, B. Behdani, and J. C. Smith, “Distributed algorithm for lifetime maximization in a delay-tolerant wireless sensor network with a mobile sink,” IEEE Transactions on Mobile Computing , vol. 12, no. 10, pp. 1920–1930, 2013

    work page 1920

  6. [4]

    Uav path planning using on-board ultrasound transducer arrays and edge support,

    E. Figetakis and A. Refaey, “Uav path planning using on-board ultrasound transducer arrays and edge support,” in 2021 IEEE International Confer- ence on Communications Workshops (ICC Workshops) , 2021, pp. 1–6

    work page 2021

  7. [5]

    Causes and effects of e-service quality for online banking,

    D.-S. Zhu and T. C.-T. Lin, “Causes and effects of e-service quality for online banking,” in 2012 13th ACIS International Conference on Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing, 2012, pp. 357–361

    work page 2012

  8. [6]

    The design and analysis of high performance online game server concurrent architecture,

    X. Wu and R. Gao, “The design and analysis of high performance online game server concurrent architecture,” in 2012 International Conference on Computer Science and Service System , 2012, pp. 1662–1665

    work page 2012

Show all 36 references
  1. [7]

    Adoption of the software-defined perimeter (sdp) architecture for infrastructure as a service,

    J. Singh, A. Refaey, and J. Koilpillai, “Adoption of the software-defined perimeter (sdp) architecture for infrastructure as a service,” Canadian Journal of Electrical and Computer Engineering , vol. 43, no. 4, pp. 357– 363, 2020

    2020

  2. [8]

    Cryptanalysis of stream cipher using density evolution,

    A. Refaey, K. Loukhaoukha, and A. Dahmane, “Cryptanalysis of stream cipher using density evolution,” in 2017 IEEE Conference on Communi- cations and Network Security (CNS) , 2017, pp. 382–383

    2017

  3. [9]

    Task sharing and scheduling for edge computing servers using hyperledger fabric blockchain,

    A. Vera-Rivera, A. Refaey, and E. Hossain, “Task sharing and scheduling for edge computing servers using hyperledger fabric blockchain,” in 2021 IEEE Globecom Workshops (GC Wkshps) , 2021, pp. 1–6

    2021

  4. [10]

    Exploring the intersection of consortium blockchain technologies and multi-access edge computing: Chronicles of a proof of concept demo,

    A. Vera-Rivera, E. Hossain, and A. R. Hussein, “Exploring the intersection of consortium blockchain technologies and multi-access edge computing: Chronicles of a proof of concept demo,” IEEE Open Journal of the Communications Society, vol. 3, pp. 2203–2236, 2022

    2022

  5. [11]

    A blockchain framework for secure task sharing in multi-access edge computing,

    A. V . Rivera, A. Refaey, and E. Hossain, “A blockchain framework for secure task sharing in multi-access edge computing,” 2020

    2020

  6. [12]

    Evolution of bitcoin and security risk in bitcoin wallets,

    P. K. Kaushal, A. Bagga, and R. Sobti, “Evolution of bitcoin and security risk in bitcoin wallets,” in 2017 International Conference on Computer, Communications and Electronics (Comptelix) , 2017, pp. 172–177

    2017

  7. [13]

    Blind detection ap- proach for ldpc, convolutional, and turbo codes in non-noisy environment,

    A. Refaey, R. Niati, X. Wang, and J. Yves-Chouinard, “Blind detection ap- proach for ldpc, convolutional, and turbo codes in non-noisy environment,” in 2014 IEEE Conference on Communications and Network Security, 2014, pp. 502–503

    2014

  8. [14]

    Multilayer authentication for communication systems based on physical-layer attributes,

    A. Refaey, W. Hou, and L. Loukhaoukha, “Multilayer authentication for communication systems based on physical-layer attributes,” Journal of Computer and Communications , vol. 2, pp. 64–75, 2014

    2014

  9. [15]

    On IoT applications: a proposed SDP framework for MQTT,

    A. Refaey, A. Sallam, and A. Shami, “On IoT applications: a proposed SDP framework for MQTT,” Electronics Letters, vol. 55, no. 22, pp. 1201– 1203, Oct. 2019

    2019

  10. [17]

    Software-defined perimeter (sdp): State of the art secure solution for modern networks,

    A. Moubayed, A. Refaey, and A. Shami, “Software-defined perimeter (sdp): State of the art secure solution for modern networks,” IEEE Network, vol. 33, no. 5, pp. 226–233, 2019

    2019

  11. [18]

    Research on consistency of distributed system based on paxos algorithm,

    W.-C. Shi and J.-P. Li, “Research on consistency of distributed system based on paxos algorithm,” in 2012 International Conference on Wavelet Active Media Technology and Information Processing (ICWAMTIP), 2012, pp. 257–259

    2012

  12. [19]

    Per- formance analysis of sdp for secure internal enterprises,

    P. Kumar, A. Moubayed, A. Refaey, A. Shami, and J. Koilpillai, “Per- formance analysis of sdp for secure internal enterprises,” in 2019 IEEE Wireless Communications and Networking Conference (WCNC), 2019, pp. 1–6

    2019

  13. [20]

    Consensus in the presence of partial synchrony,

    C. Dwork, N. Lynch, and L. Stockmeyer, “Consensus in the presence of partial synchrony,” J. ACM , vol. 35, no. 2, p. 288–323, Apr. 1988. [Online]. Available: https://doi.org/10.1145/42282.42283 8

    1988

  14. [21]

    Securing smart home networks with software-defined perimeter,

    A. Sallam, A. Refaey, and A. Shami, “Securing smart home networks with software-defined perimeter,” in 2019 15th International Wireless Communications & Mobile Computing Conference (IWCMC) , 2019, pp. 1989–1993

    2019

  15. [22]

    Blockchain-enabled smart contracts: Architecture, applications, and fu- ture trends,

    S. Wang, L. Ouyang, Y . Yuan, X. Ni, X. Han, and F.-Y . Wang, “Blockchain-enabled smart contracts: Architecture, applications, and fu- ture trends,” IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 49, no. 11, pp. 2266–2277, 2019

    2019

  16. [23]

    Evaluation and demon- stration of blockchain applicability framework,

    S. N. G. Gourisetti, M. Mylrea, and H. Patangia, “Evaluation and demon- stration of blockchain applicability framework,” IEEE Transactions on Engineering Management, vol. 67, no. 4, pp. 1142–1156, 2020

    2020

  17. [25]

    Localization for in- telligent systems using unsupervised learning and prediction approaches,

    P. Mirdita, Z. Khaliq, A. R. Hussein, and X. Wang, “Localization for in- telligent systems using unsupervised learning and prediction approaches,” IEEE Canadian Journal of Electrical and Computer Engineering , vol. 44, no. 4, pp. 443–455, 2021

    2021

  18. [26]

    When blockchain meets sgx: An overview, challenges, and open issues,

    Z. Bao, Q. Wang, W. Shi, L. Wang, H. Lei, and B. Chen, “When blockchain meets sgx: An overview, challenges, and open issues,” IEEE Access, vol. 8, pp. 170 404–170 420, 2020

    2020

  19. [27]

    On scaling decentralized blockchains - (a position paper),

    K. Croman, C. Decker, I. Eyal, A. E. Gencer, A. Juels, A. Kosba, A. Miller, P. Saxena, E. Shi, E. G. Sirer, D. Song, and R. Wattenhofer, “On scaling decentralized blockchains - (a position paper),” in Financial Cryptography Workshops, 2016

    2016

  20. [28]

    Computation offloading and content caching in wireless blockchain networks with mobile edge computing,

    M. Liu, F. R. Yu, Y . Teng, V . C. M. Leung, and M. Song, “Computation offloading and content caching in wireless blockchain networks with mobile edge computing,” IEEE Transactions on Vehicular Technology , vol. 67, no. 11, pp. 11 008–11 021, 2018

    2018

  21. [29]

    Ai soc-based accelerator for speech classification ai soc-based accelerator for speech classification,

    C. DeSantis and A. R. Hussein, “Ai soc-based accelerator for speech classification ai soc-based accelerator for speech classification,” IEEE Canadian Journal of Electrical and Computer Engineering , vol. 45, no. 3, pp. 222–231, 2022

    2022

  22. [30]

    An intelligent transaction migration scheme for raft-based private blockchain in internet of things applications,

    H. Lu, X. Xu, K. Zheng, and X. Wang, “An intelligent transaction migration scheme for raft-based private blockchain in internet of things applications,” IEEE Communications Letters , vol. PP, pp. 1–1, 05 2021

    2021

  23. [31]

    Performance analysis of the raft con- sensus algorithm for private blockchains,

    D. Huang, X. Ma, and S. Zhang, “Performance analysis of the raft con- sensus algorithm for private blockchains,” IEEE Transactions on Systems, Man, and Cybernetics: Systems , vol. 50, no. 1, pp. 172–181, 2020

    2020

  24. [32]

    A survey of decentralizing applications via blockchain: The 5g and beyond perspective,

    K. Yue, Y . Zhang, Y . Chen, Y . Li, L. Zhao, C. Rong, and L. Chen, “A survey of decentralizing applications via blockchain: The 5g and beyond perspective,” IEEE Communications Surveys & Tutorials , vol. 23, no. 4, pp. 2191–2217, 2021

    2021

  25. [33]

    Evolved prevention strategies for 6g networks through stochastic games and reinforcement learning,

    E. Figetakis, A. R. Hussein, and M. Ulema, “Evolved prevention strategies for 6g networks through stochastic games and reinforcement learning,” IEEE Networking Letters , vol. 5, no. 3, pp. 164–168, 2023

    2023

  26. [34]

    Autonomous mec selection in federated next- gen networks via deep reinforcement learning,

    E. Figetakis and A. Refaey, “Autonomous mec selection in federated next- gen networks via deep reinforcement learning,” in GLOBECOM 2023 - 2023 IEEE Global Communications Conference , 2023, pp. 2045–2050. Algorithm 1: DDPG-based Multi-access Edge Comput- ing Resource Training ...

    2023

  27. [35]

    Initialization: Set iteration counter Γ = 0

  28. [36]

    Update Γ and S: Γ ← Γ + 1, S ← SΓ+1 if Γ > γthen Sample a mini-batch e from δ

    Training Loop: while Γ ̸= Γmax do Sample a continuous action: A = π(S | ψ) + g, g ∼ G Update action refinement: A∗ = arg max ˜A∈A ˆQ(S, ˜A) Execute action A, observe reward R, and transition to the next state: SΓ+1 = f (S, A) Store the observation (S, A, R, SΓ+1) in δ. Update ...

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.