Pith. sign in

REVIEW 3 major objections 5 minor 26 references

Pareto-Aware Hierarchical Reinforcement Learning for Online Resource Allocation in RIS-assisted Large-Scale IoT Systems

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read This paper claims that every efficient rate allocation in a RIS-assisted IoT network can be encoded by a direction vector, allowing reinforcement learning to act on a tiny space instead of thousands of beamforming and phase variables.

desk verdict The directional Pareto mapping that carries the paper is not a bijection, so the lossless-compression claim breaks; the engineering is a repairable idea but the paper's central theorem is false as stated. read the letter →

arxiv 2608.13032 v1 pith:N3BFO7QT submitted 2026-08-13 cs.NI

Wenhan Xu , Jiashuo Jiang , Danny H.K. Tsang This is my paper · ORCID
classification cs.NI
keywords ReconfigurableintelligentsurfaceHierarchicalreinforcementlearningParetooptimalityDirectionalmax-minscalingAutoencodercompressionOnlineresourceallocationMobileedgecomputingInternetofThings
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 proposes a way to make reinforcement learning practical for real-time resource allocation in RIS-assisted IoT networks by cutting the action space in two stages. The first stage replaces the high-dimensional physical variables—transmit beamformers, RIS phase shifts, and reflection amplitudes—with a low-dimensional weight vector $\omega$ using a directional max-min scaling problem. The paper argues this mapping is lossless: every Pareto-optimal rate vector of the achievable rate region, convex or not, is reachable by some direction $\omega$, so no efficient operating point is lost. The second stage compresses $\omega$ further with an offline-trained autoencoder, so the RL agent acts on a tiny latent vector while a frozen decoder and the model-driven projection restore physical configurations. Simulations on a multi-user mobile edge computing network indicate faster convergence, shorter training time, and lower delay-plus-energy cost than weighted-sum-rate or uncompressed RL baselines.

What carries the argument

The load-bearing object is the directional max-min scaling problem P3. It replaces linear scalarization (weighted sum-rate) with a search along a ray: for a direction vector $\omega$, maximize a scalar $R_s$ such that each user's rate is at least $\omega_k R_s$; the outer loop is a binary search on $R_s$, and the inner loop is a feasibility problem P4 solved by alternating optimization with successive convex approximation, producing convex subproblems for beamforming and for RIS coefficients. The second mechanism is a symmetric autoencoder whose frozen decoder $D_\psi$ is embedded in the online loop, translating the RL agent's low-dimensional latent action $Z$ into the weight-and-application vector $(\omega, X)$ that P3 then maps to a Pareto-optimal physical configuration.

What would settle it

Take a two-user rate region that is the union of the downsets of $(1,10)$ and $(10,1)$. This region is non-convex and contains $(1,1)$, but the point $(1,10)$ dominates $(1,1)$. Solving P3 with direction $\omega=(1,1)$ returns $R_s=1$ and the rate vector $(1,1)$, a dominated point, even though a Pareto point exists with the same first coordinate. This directly contradicts the claimed one-to-one correspondence and can be checked with the paper's own feasibility solver on two users.

Watch

Extended reading notes

Core claim

The central claim is that the set of Pareto-optimal operating points of an RIS-assisted multi-user network can be indexed bijectively by normalized direction vectors in the positive orthant, and that the directional max-min scaling problem P3—maximize a common scaling factor $R_s$ subject to $R_k \ge \omega_k R_s$ for every user—recovers the unique Pareto-optimal rate vector for a chosen direction $\omega$. Because any sensible objective is non-decreasing in the user rates, its global optimum lies on this frontier, so optimizing over $\omega$ and application-level variables $X$ is equivalent to optimizing over the original $\{f,\beta,\theta,X\}$. The paper then chains an autoencoder onto this mapping, so the RL agent outputs only a latent code $Z$; the decoder reconstructs the weight vector, and P3 projects it onto the Pareto boundary. The claimed payoff is that PAAERL inherits the geometric exactness of P3 while avoiding both action-space explosion and the information loss the paper attributes to weighted sum-rate scalarization in non-convex rate regions.

Load-bearing premise

The load-bearing premise is that each direction vector maps to exactly one Pareto-optimal rate point, so replacing the physical variables with $\omega$ loses no information even when the rate region is non-convex.

Editorial extensions

If this is right

  • The number of RIS elements $M$ stops being part of the RL action space: once the P3 projection is in place, the policy sees only $K$ weight coordinates plus application variables, regardless of $M$.
  • Any network objective that is non-decreasing in user rates—latency, energy, sum rate—can be optimized by searching the same low-dimensional weight space, so the architecture transfers to STAR-RIS, UAV-relayed, and other coupled RIS systems.
  • Because the autoencoder is trained offline on Pareto-optimal profiles, online execution is a single decoder pass plus one P3 projection per slot, making decisions fast enough for quasi-static channels even in dense deployments.
  • The paper's convergence plots show the Pareto-aware variants reach lower steady-state cost than weighted-sum-rate variants, which the paper reads as evidence that the non-convex 'dented' parts of the rate region are exactly where P3 helps.

Reading between the lines

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

  • If the one-to-one claim fails, PAAERL would degrade gracefully into an approximate compressor whose error is the depth of the dent in the rate region; the paper's own caveat about autoencoder reconstruction error already admits approximate behavior in the second stage.
  • A minimal two-user test of the geometric claim is to run P3 for direction $(1,1)$ on a rate region formed by the union of the downsets of $(1,10)$ and $(10,1)$ and check whether the returned point is dominated.
  • If the geometric mapping is only approximately bijective in practice, the autoencoder could be trained to avoid directions that point into dents, effectively learning which rays to distrust; this is a testable extension the paper does not explore.
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

3 major / 5 minor

Summary. The manuscript proposes PAAERL, a hierarchical reinforcement-learning framework for online resource allocation in RIS-assisted multi-user IoT/MEC systems. The first compression stage replaces the high-dimensional physical variables (transmit beamforming, RIS phase shifts and amplitudes) with a positive weight vector ω_PT by solving a directional max-min scaling problem P3, and claims that this establishes a bijective, information-lossless mapping between weight vectors and the Pareto-optimal frontier of the achievable rate region, regardless of convexity. A second, data-driven autoencoder compresses the priority profile into a low-dimensional latent action space for a PPO agent. The paper provides Algorithms 1 and 2, a complexity analysis, and simulations in an RIS-assisted MEC scenario comparing six strategies on convergence speed, total cost, and training time.

Significance. If the bijective Pareto-representation claim were valid, the proposed model-driven plus data-driven compression would be a conceptually attractive way to decouple physical-layer optimization from high-level control, and the simulation results suggest practical gains in training time and cost. The paper is also transparent that the autoencoder stage is lossy, which is a reasonable design choice. However, the central theoretical guarantee—the claimed lossless substitution of the physical variables by the weight vector—is false as stated. The directional max-min problem P3 is not a single-valued map onto the Pareto frontier for non-convex comprehensive rate regions, and the inner feasibility solver is not shown to be an exact oracle. Because the hierarchical RL design and the paper's central 'Pareto-aware' claim rest on this guarantee, the main contribution is not established in the submitted form.

major comments (3)
  1. [Section III.C, P3 (Eq. (11))] The claimed one-to-one correspondence between positive-orthant direction vectors and Pareto-optimal points is false for non-convex comprehensive rate regions. For R = ([0,2]×[0,1]) ∪ ([0,1]×[0,2]) and ω=(1,1), P3 gives max R_s = 1, but the feasible rate set at R_s=1 contains the dominated point (1,1) as well as the Pareto points (2,1) and (1,2); hence the ray in direction ω does not terminate at a Pareto point, and P3 does not define a unique rate vector. The observation that the ray through a given Pareto point is unique does not imply that every ray contains a Pareto point. This invalidates the deterministic mapping M:ω_PT→R* and the 'no theoretical information loss' claim that follows it.
  2. [Section III.C, P3 constraints] Even apart from the counterexample, the constraints in P3 are R_k ≥ ω_k R_s, not equalities. Consequently, for a fixed optimal R_s there are generally many feasible rate vectors satisfying the constraints, so the output of P3 depends on the arbitrary choice of the inner feasibility solver; the paper never specifies a tie-breaking rule. A directional representation of the Pareto frontier requires either equality constraints or an additional projection step that selects a definite rate vector. The existence result that can actually be proved is weaker: for every Pareto point p, choosing ω=p and solving P3 returns p (if the feasibility oracle is exact). The paper should replace the bijection theorem with this statement and re-derive the RL justification from it.
  3. [Section III.D, P5 and Algorithm 1] Algorithm 1's binary search uses P4/P5 as an exact feasibility oracle, but P5 is not validated as one. P5 minimizes the sum of slack variables d_k without imposing d_k ≥ 0, so a zero-slack feasible point need not be the minimizer. More importantly, P5a/P5b replace the nonconvex signal-power term by an SCA lower bound; this relaxes the constraint, and the alternating SCA procedure is only claimed to reach a stationary point of the relaxed problem. A stationary point can have positive slack even when P4 is feasible, causing the bisection to misclassify a feasible R_s as infeasible. The paper provides no convergence or exactness proof showing that the binary search returns the true maximum R_s; the statement that the 'SCA lower bounds guarantee that the objective value is non-decreasing and bounded' is not by itself sufficient for a minimization problem.
minor comments (5)
  1. [Section II and Eq. (6)] Please clean up the notation typos: the reference to Table I appears as 'Table.', and Eq. (6) has bracket/brace errors around the logarithm expression.
  2. [Section III.C] The vector ω_PT is introduced before its normalization or domain is defined; specify whether it is normalized to sum to one, lies on the unit sphere, or is otherwise made scale-invariant, and explain how scaling redundancy is handled.
  3. [Algorithm 1, line 11] The instruction to 'find R_max when P4 has no feasible solution' is ambiguous; the bisection initialization and the convention for checking feasibility at R_s=0 should be stated precisely.
  4. [Section III.E] The statement that imperfect AE reconstruction 'will only cause a small amount of performance loss' is not quantified; the simulations do not report reconstruction error or its effect on the achieved cost, so this claim is currently unsupported.
  5. [Section IV] The convergence and scaling plots do not include error bars or multiple-seed statistics, so the reported improvements over WSRRL and PARL cannot be separated from run-to-run variance.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the P3 directional mapping is an analytic construction, the autoencoder is explicitly approximate, and cited prior work is background rather than load-bearing.

full rationale

The paper's central derivation, the directional max-min mapping M: omega^PT -> R*, is defined by solving problem P3 (Section III.C) and is not fitted to data, not defined in terms of the RL objective, and not imported from the authors' previous work. The claim that every Pareto point corresponds to a positive-orthant direction is an analytic geometric assertion, and the autoencoder phase (Section III.E) is openly lossy: the paper states that 'If AE cannot perfectly replace parameters, it will only cause a small amount of performance loss,' so no hidden circularity is introduced by training the AE on data produced by the same P3 mapping. References [18]-[20] are cited as prior contributions to similar RIS optimization problems, not as the source of the bijectivity or losslessness guarantee, so the self-citations are not load-bearing. The main theoretical vulnerability is correctness, not circularity: Section III.C's statement that uniqueness of the ray through a Pareto point 'implies a one-to-one correspondence between the set of Pareto optimal points and the set of direction vectors in the positive orthant' is mathematically unproven, and P3's inequality constraints R_k >= omega_k^PT R_s can admit multiple rate vectors for a fixed direction, especially in non-convex comprehensive rate regions. That is an invalid or incomplete proof of the claimed guarantee, but it is not a reduction of the result to its own inputs. Since the derivation chain does not fit parameters and rename them as predictions, and does not rely on self-citation for its central equivalence, there is no significant circularity.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The central theoretical claim rests on a geometric assumption about the rate region that is not generally true, and on a convergence assumption for the inner feasibility solver. No new physical entities are introduced; the autoencoder is a standard data-driven component whose hyperparameters are free. The paper's own admission that AE reconstruction causes performance loss in the RIS-element scaling experiment shows the fitted AE is essential to the headline speedups.

free parameters (2)
  • Latent code dimension d = not specified
    The autoencoder's bottleneck dimension d determines how much the priority space is compressed; it is a hand-chosen hyperparameter that directly controls the reconstruction error and the final network cost. The paper does not report d or a sensitivity study.
  • Autoencoder architecture and training hyperparameters = not specified
    Number of layers, neurons, learning rate, and training epochs for the AE are not given; these fitted weights affect the decoder's reconstruction fidelity and thus the PAAERL performance.
assumptions (3)
  • domain assumption The achievable rate region R is such that every ray from the origin intersects the boundary at a single point that is Pareto-optimal.
    Invoked in Section III.C to claim one-to-one correspondence; false for comprehensive non-convex regions.
  • domain assumption The AO-SCA algorithm for P5 converges to a feasible point with zero slack variables d_k for the original feasibility problem P4.
    Section III.D states convergence to a stationary point of the relaxed problem; the paper does not prove that the stationary point satisfies the original SINR constraints, which is required for the binary search over R_s to be valid.
  • domain assumption The application-specific objective is a non-decreasing function of user rates, so the global optimum lies on the Pareto frontier.
    Section III.A argues that all efficient points are on the Pareto boundary; this is standard and reasonable.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Pareto-Aware Hierarchical Reinforcement Learning for Online Resource Allocation in RIS-assisted Large-Scale IoT Systems." pith.science (2026). https://pith.science/paper/N3BFO7QT

@misc{pith2026260813032,
  author       = {Pith},
  title        = {Pith review of: Pareto-Aware Hierarchical Reinforcement Learning for Online Resource Allocation in RIS-assisted Large-Scale IoT Systems},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N3BFO7QT}},
  note         = {Machine review of arXiv:2608.13032}
}
read the original abstract

With the rapid evolution of 5G and emerging 6G networks, reconfigurable intelligent surfaces (RIS) have become a critical technology for enhancing wireless communication scenarios. However, optimizing RIS-assisted multi-user systems typically introduces high-dimensional physical layer variables and non-convex Pareto-optimal rate sets, posing severe computational challenges for real-time applications. To address these limitations, this paper proposes a dimension-reduced, hierarchical reinforcement learning (RL) framework, termed Pareto-aware autoencoder-assisted RL (PAAERL), to optimize online resource allocation in RIS-assisted Internet of Things (IoT) networks. Our approach first substitutes high-dimensional continuous RIS beamforming variables with lower-dimensional weight vectors that strictly represent the Pareto-optimal frontier, theoretically avoiding geometric information loss across both convex and non-convex rate regions. To further mitigate the curse of dimensionality in dense networks, an autoencoder architecture is integrated to execute a secondary, data-driven compression phase, mapping the priority space into a highly condensed continuous latent action space. Extensive simulations conducted across practical communication scenarios, including multi-user mobile edge computing (MEC) networks, demonstrate that the proposed PAAERL framework drastically reduces offline training times, accelerates online policy convergence, and significantly decreases overall network costs compared to state-of-the-art benchmarks, underscoring its exceptional scalability and practical viability for next-generation intelligent IoT environments.

Figures

Figures reproduced from arXiv: 2608.13032 by the authors.

Figure 1
Figure 1. The wireless communication model of an uplink RIS-aided IoT [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. The structure of autoencoder. pose challenges in massive IoT networks where K is large. Furthermore, spatial correlation among user distributions and recurring environmental dynamics often introduce statistical redundancies within the priority vector space. To exploit these latent statistical regularities and achieve an ultra-low￾dimensional action representation for the RL agent, we intro￾duce a data-driven seconda… view at source ↗
Figure 4
Figure 4. The total cost C vs. number of users of our proposed PAAERL and benchmarks. reward pathways. In contrast, algorithms embedded with data￾driven action reduction, such as AERL, WSRAERL, and the proposed PAAERL, exhibit steep learning curves in the early training phases from 0 to 150 iterations. By utilizing the deep autoencoder to map policies onto an ultra-low-dimensional latent continuous code Z, the agent circumven… view at source ↗
Figures from the paper (2 more)
Figure 3
Figure 3. Figure 3: The convergence behavior of our proposed PAAERL with benchmarks. [PITH_FULL_IMAGE:figures/full_fig_p010_3.png]
Figure 5
Figure 5. Figure 5: The total cost C vs. number of RIS elements of our proposed PAAERL and benchmarks. users, scaling K triggers a compound explosion in the dimen￾sionality of the action space. The conventional model-driven compression benchmarks, namely WSRRL and WSRAERL, achieve lower n…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 23 canonical work pages

  1. [1]

    Multi-objective optimization of 3-D cell deployment in sustainable B5G/6G networks: Balancing perfor- mance and sustainability,

    W.-C. Chien, G. Jeon, and H.-H. Cho, “Multi-objective optimization of 3-D cell deployment in sustainable B5G/6G networks: Balancing perfor- mance and sustainability,”IEEE Trans. Network Service Management, vol. 22, no. 4, pp. 3077–3091, 2025

    work page 2025

  2. [2]

    Graph neural networks for quality of service improvement in interference-limited ultra-reliable and low-latency communications,

    Y . Liu, C. She, Y . Zhong, W. Hardjawana, F.-C. Zheng, and B. Vucetic, “Graph neural networks for quality of service improvement in interference-limited ultra-reliable and low-latency communications,” IEEE Trans. Vehicular Technol., vol. 73, no. 3, pp. 3718–3732, 2023

    work page 2023

  3. [3]

    Multi-dimensional QoS evaluation and optimization of mobile edge computing for IoT: A survey,

    J. Huang, F. Liu, and J. Zhang, “Multi-dimensional QoS evaluation and optimization of mobile edge computing for IoT: A survey,”Chinese J. Electronics, vol. 33, no. 4, pp. 859–874, 2024

    work page 2024

  4. [4]

    Energy-efficient trans- mission strategy for UA V-RIS 2.0 assisted communications using rate splitting multiple access,

    A. M. Huroon, Y .-C. Huang, and L.-C. Wang, “Energy-efficient trans- mission strategy for UA V-RIS 2.0 assisted communications using rate splitting multiple access,”IEEE Trans. Wireless Commun., vol. 25, pp. 5246–5261, 2026

    work page 2026

  5. [5]

    Optimization of secure computation efficiency in UA V-enabled RIS- assisted MEC-IoT networks with aerial and ground eavesdroppers,

    E. T. Michailidis, M.-G. V olakaki, N. I. Miridakis, and D. V ouyioukas, “Optimization of secure computation efficiency in UA V-enabled RIS- assisted MEC-IoT networks with aerial and ground eavesdroppers,” IEEE Trans. Commun., vol. 72, no. 7, pp. 3994–4009, 2024

    work page 2024

  6. [6]

    Reconfigurable intelligent surfaces: Principles and opportunities,

    Y . Liu, X. Liu, X. Mu, T. Hou, J. Xu, M. Di Renzo, and N. Al-Dhahir, “Reconfigurable intelligent surfaces: Principles and opportunities,”IEEE communications surveys & tutorials, vol. 23, no. 3, pp. 1546–1577, 2021

    2021

  7. [7]

    Reconfigurable intelligent surfaces for 6G: Emerging hardware architectures, applications, and open challenges,

    E. Basar, G. C. Alexandropoulos, Y . Liu, Q. Wu, S. Jin, C. Yuen, O. A. Dobre, and R. Schober, “Reconfigurable intelligent surfaces for 6G: Emerging hardware architectures, applications, and open challenges,” IEEE Vehicular Technol. Magaz., vol. 19, no. 3, pp. 27–47, 2024

    work page 2024

  8. [8]

    Interference nulling using reconfigurable intelligent surface,

    T. Jiang and W. Yu, “Interference nulling using reconfigurable intelligent surface,”IEEE J. Selected Areas Commun., vol. 40, no. 5, pp. 1392– 1406, 2022

    work page 2022

Show all 26 references
  1. [9]

    Double self- sustainable reconfigurable intelligent surfaces aided wireless communi- cations,

    J. Wang, S. Luo, Y . Li, W. Xie, X. Li, and A. Nallanathan, “Double self- sustainable reconfigurable intelligent surfaces aided wireless communi- cations,”IEEE Trans. Vehicular Technol., vol. 73, no. 11, pp. 17 869– 17 874, 2024

    2024

  2. [10]

    Weighted sum-rate maximization by joint antenna grouping and movable RIS deployment,

    J. Tang, Z. Lv, J. Xiao, J. Wu, and B. Shim, “Weighted sum-rate maximization by joint antenna grouping and movable RIS deployment,” IEEE Trans. Wireless Commun., 2025

    2025

  3. [11]

    Achievable rate region maxi- mization in intelligent reflecting surfaces-assisted interference channel,

    M. Jiang, Y . Li, G. Zhang, and M. Cui, “Achievable rate region maxi- mization in intelligent reflecting surfaces-assisted interference channel,” IEEE Trans. Vehicular Technol., vol. 70, no. 12, pp. 13 406–13 412, 2021

    2021

  4. [12]

    Achievable rate region of MISO interference channel aided by intelligent reflecting surface,

    W. Huang, Y . Zeng, and Y . Huang, “Achievable rate region of MISO interference channel aided by intelligent reflecting surface,”IEEE Trans. Vehicular Technol., vol. 69, no. 12, pp. 16 264–16 269, 2020

    2020

  5. [13]

    Energy harvesting reconfigurable intelligent surface for UA V based on robust deep reinforcement learning,

    H. Peng and L.-C. Wang, “Energy harvesting reconfigurable intelligent surface for UA V based on robust deep reinforcement learning,”IEEE Trans. Wireless Commun., vol. 22, no. 10, pp. 6826–6838, 2023

    2023

  6. [14]

    Energy efficiency optimization for IoT systems with reconfigurable intelligent surfaces: A self-supervised reinforcement learning approach,

    H. Peng, Y .-T. Lin, C.-Y . Ho, and L.-C. Wang, “Energy efficiency optimization for IoT systems with reconfigurable intelligent surfaces: A self-supervised reinforcement learning approach,”IEEE Trans. Wireless Commun., 2025

    2025

  7. [15]

    Deep reinforcement learning for practical phase-shift optimization in RIS- aided MISO URLLC systems,

    R. Hashemi, S. Ali, N. H. Mahmood, and M. Latva-Aho, “Deep reinforcement learning for practical phase-shift optimization in RIS- aided MISO URLLC systems,”IEEE Internet Things J., vol. 10, no. 10, pp. 8931–8943, 2022

    2022

  8. [16]

    Learning to focus: CSI-free hierarchical MARL for reconfigurable reflectors,

    H. Le, M. Ibrahim, O. Bedir, J. Tao, and S. Ekin, “Learning to focus: CSI-free hierarchical MARL for reconfigurable reflectors,”arXiv preprint arXiv:2604.05165, 2026

    2026 arXiv

  9. [17]

    Multiple reconfigurable intelli- gent surfaces aided V2X offloading networks: A federated reinforcement learning-based approach,

    M. Zeng, Y . Zhao, J. Wang, and Z. Fei, “Multiple reconfigurable intelli- gent surfaces aided V2X offloading networks: A federated reinforcement learning-based approach,”IEEE Trans. Vehicular Technol., vol. 75, no. 1, pp. 1281–1294, 2026

    2026

  10. [18]

    Joint channel estima- tion and reinforcement-learning-based resource allocation of intelligent- reflecting-surface-aided multicell mobile edge computing,

    W. Xu, J. Yu, Y . Wu, and D. H. K. Tsang, “Joint channel estima- tion and reinforcement-learning-based resource allocation of intelligent- reflecting-surface-aided multicell mobile edge computing,”IEEE Inter- net Things J., vol. 11, no. 7, pp. 11 862–11 875, 2023

    2023

  11. [19]

    Reconfigurable intelligent surface aided mobile fog computing: A space aggregation- based Lyapunov driven reinforcement learning approach,

    W. Xu, C. Pan, Y . Yuan, Y . Wu, and D. H. K. Tsang, “Reconfigurable intelligent surface aided mobile fog computing: A space aggregation- based Lyapunov driven reinforcement learning approach,”IEEE Trans. Mobile Comput., vol. 25, no. 3, pp. 3809–3823, 2026

    2026

  12. [20]

    Enhancing fog computing through intelligent reflecting surface assistance: A Lyapunov driven reinforcement learning approach,

    W. Xu, Y . Yuan, and D. H. K. Tsang, “Enhancing fog computing through intelligent reflecting surface assistance: A Lyapunov driven reinforcement learning approach,” in2024 IEEE 10th World Forum on Internet of Things (WF-IoT). IEEE, 2024, pp. 759–764

    2024

  13. [21]

    Weighted sum- rate maximization for reconfigurable intelligent surface aided wireless networks,

    H. Guo, Y .-C. Liang, J. Chen, and E. G. Larsson, “Weighted sum- rate maximization for reconfigurable intelligent surface aided wireless networks,”IEEE Trans. Wireless Commun., vol. 19, no. 5, pp. 3064– 3076, 2020

    2020

  14. [22]

    Autoencoder- based communications with reconfigurable intelligent surfaces,

    T. Erpek, Y . E. Sagduyu, A. Alkhateeb, and A. Yener, “Autoencoder- based communications with reconfigurable intelligent surfaces,” in2021 IEEE Internation. Symposium Dynamic Spectrum Access Networks (DySPAN). IEEE, 2021, pp. 242–247

    2021

  15. [23]

    Aerial intelligent reflecting surface: Joint placement and passive beamforming design with 3D beam flattening,

    H. Lu, Y . Zeng, S. Jin, and R. Zhang, “Aerial intelligent reflecting surface: Joint placement and passive beamforming design with 3D beam flattening,”IEEE Trans. Wireless Commun., vol. 20, no. 7, pp. 4128– 4143, 2021

    2021

  16. [24]

    Lyapunov optimization based mobile edge computing for internet of vehicles systems,

    Y . Jia, C. Zhang, Y . Huang, and W. Zhang, “Lyapunov optimization based mobile edge computing for internet of vehicles systems,”IEEE Trans. Commun., vol. 70, no. 11, pp. 7418–7433, 2022

    2022

  17. [25]

    Lyapunov-guided deep reinforcement learning for stable online computation offloading in mobile-edge computing networks,

    S. Bi, L. Huang, H. Wang, and Y .-J. A. Zhang, “Lyapunov-guided deep reinforcement learning for stable online computation offloading in mobile-edge computing networks,”IEEE Trans. Wireless Commun., vol. 20, no. 11, pp. 7519–7537, 2021

    2021

  18. [26]

    Dynamic mobile edge computing empowered by reconfigurable intelligent surfaces,

    P. Di Lorenzo, M. Merluzzi, and E. C. Strinati, “Dynamic mobile edge computing empowered by reconfigurable intelligent surfaces,” in2021 IEEE 22nd International Workshop Signal Process. Advances Wireless Commun. (SPAWC). IEEE, 2021, pp. 526–530

    2021

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.