Pith. sign in

REVIEW 5 major objections 5 minor 44 references

Percentile-Based Deep Reinforcement Learning and Reward Based Personalization For Delay Aware RAN Slicing in O-RAN

T0 review · 5 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A percentile-based DRL policy meets RAN delay bounds within 1% while cutting average delay by 38%; a reward-weighted model-sharing method beats federated averaging across ten environments.

desk verdict A serious DRL-for-RAN-slicing paper with a genuine reward-shaping idea and a clean personalization scheme, but the headline numbers rest on undisclosed coefficients and single-run statistics; worth refereeing, not desk-rejecting. read the letter →

arxiv 2507.18111 v1 pith:VMUA6A5I submitted 2025-07-24 cs.LG

Peyman Tehrani , Anas Alsoliman This is my paper · ORCID
classification cs.LG
keywords RANslicingO-RANdeepreinforcementlearningpercentile-baseddelayguaranteerewardshapingpersonalizedfederatedPRBallocationprobabilisticQoS
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 takes aim at a specific failure mode in deep reinforcement learning (DRL) for radio access network (RAN) slicing: optimizing for average delay does not meet percentile-style quality-of-service (QoS) guarantees. The authors derive a reward function from the law of large numbers that rewards the probability of meeting a per-packet delay deadline, then reshape it into a practical, clipped form that avoids an 'unlearnable region' where the theoretical reward gives no usable gradient. Their agent, PDA-DRL, is claimed to satisfy the probabilistic delay constraint within a 1 percent margin while using roughly the same number of physical resource blocks (PRBs) as a mean-delay-optimized baseline, yet delivering 38 percent lower average delay and 33 percent lower delay standard deviation. The paper also proposes a reward-based personalization method that lets multiple MVNOs share DRL model weights by weighting each other's models by measured performance, which it shows beats federated averaging and similarity-based aggregation across ten environments.

What carries the argument

The load-bearing object is the practical reward function of Eq. (11): $r = -\Delta \gamma_p + \exp(\zeta_p \Delta + \nu_p N_{Ts}^2)$ when $\Delta \geq 0$ and $r = \Delta \gamma_n + \exp(\zeta_n \Delta + \nu_n N_{Ts})$ when $\Delta < 0$, with $\Delta = Pr(d_q < D_{max}) - (1-\epsilon)$ and the output clipped to $(-R_{max}, 0)$. This shape is designed so that when the satisfaction probability is below target the exponential term grows, forcing the agent to request more PRBs quickly, while near the feasible boundary the linear term in $\Delta$ dominates and the PRB penalty $N_{Ts}$ prevents over-allocation. The second object is the personalization coefficient of Eq. (16): $\alpha_{i,j} = \exp(\beta \hat{R}^T_{i,j}) / \sum_j \exp(\beta \hat{R}^T_{i,j})$, where $\hat{R}^T_{i,j}$ is the average reward that agent $i$ obtains by running agent $j$'s model on its own environment for $T$ episodes; $\beta$ interpolates between plain averaging ($\beta \approx 0$) and hard selection of the best model ($\beta$ large). Both objects carry the paper's arguments: the reward function converts a hard probabilistic constraint into a smooth learning signal, and the personalization rule converts measured performance into aggregation weights.

What would settle it

Run PDA-DRL in an environment whose satisfaction-probability curve rises slowly or plateaus before the target $(1-\epsilon)$ is reached, and compare the PRB count at the shaped reward's maximum with the Lagrangian optimum. If the piecewise reward's argmax shifts (or the clipping $R_{max}$ flattens the landscape), the 1% QoS margin or the 38% delay reduction would not reproduce; the paper reports only one such comparison in Fig. 3a.

Watch

Extended reading notes

Core claim

The central claim is that a percentile-based, reward-shaped DRL controller can satisfy a probabilistic upper bound on packet transmission delay in O-RAN slicing while spending nearly the same radio resources as a baseline that only minimizes average delay. The reward is built from the empirical satisfaction probability $Pr(d_q < D_{max})$ minus the target $(1-\epsilon)$, multiplied by a trade-off parameter and balanced against PRB usage, which the authors show is equivalent to the Lagrangian dual of the constrained minimization problem. Because the theoretical reward has a flat 'unlearnable region' where no packets meet the deadline, the authors replace it with a piecewise exponential/linear shape whose maximum coincides with the theoretical optimum (shown for one scenario) and clip it to $(-R_{max}, 0)$. In simulation, the resulting PDA-DRL policy meets the QoS constraint within a 1% margin, cuts average delay by 38% and delay standard deviation by 33% versus a mean-delay DRL baseline, and uses about the same PRB count. For model sharing, the paper's reward-based personalization, where each agent weighs other agents' weights by $\exp(\beta \times \text{average reward})$ on its own environment, outperforms federated averaging and both feature- and weight-similarity aggregation across ten distinct environments.

Load-bearing premise

The load-bearing premise is that the practical, clipped reward function in Eq. (11) reaches its maximum at the same number of PRBs as the idealized Lagrangian reward for every environment, so that reshaping and clipping do not move the optimum; the paper only illustrates this for a single scenario and gives no coefficient values or tuning procedure for $\zeta, \nu, \gamma, \lambda, R_{max}$.

Editorial extensions

If this is right

  • Satisfying a probabilistic delay bound with DRL is achievable with a reward that directly tracks the empirical satisfaction probability, not just the mean delay.
  • The shaped reward gives a principled way to escape a failure mode (zero gradient when no packet meets the deadline) that plagues naive Lagrangian rewards in constraint-heavy wireless tasks.
  • Reward-based personalization can replace federated averaging when agents operate in very different environments, because each agent evaluates candidate models on its own soil.
  • Since PRB usage stays nearly constant while delay variance shrinks, the approach promises more predictable latency for time-critical slices without extra radio resources.

Reading between the lines

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

  • The reward-shaping trick could be applied to any DRL problem with a hard threshold constraint and a failure region where the naive reward is flat, such as power or budget limits in other scheduling domains.
  • A practical tuning recipe for the coefficients $\zeta, \nu, \gamma, \lambda, R_{max}$ (e.g., a grid search or a schedule that anneals them) would be needed before the method can be deployed without environment-specific hand-tuning.
  • The personalization rule's reliance on testing other agents' models on one's own environment assumes a digital twin or emulator is available; in a purely live network that evaluation cost may be prohibitive.
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 addresses PRB allocation for RAN slicing in O-RAN under probabilistic delay constraints. It derives a Lagrangian-style reward from the constrained optimization problem (Eq. 10), then proposes a shaped percentile-based reward (Eq. 11) and trains a policy-gradient agent (PDA-DRL) that minimizes PRB usage while keeping Pr(d < Dmax) close to 1-epsilon. It also proposes a reward-based personalization scheme (Eq. 16) in which each MVNO aggregates other agents' model weights according to their measured performance in its own environment. Simulation results report that PDA-DRL meets the QoS constraint within a 1% margin while reducing average delay by 38% and delay STD by 33% over an average-delay DRL baseline, and that reward-based personalization outperforms federated averaging and similarity-based aggregation across 10 environments. Additional experiments on the Colosseum/SCOPE testbed are described in Appendix C.

Significance. If the claims hold, the paper is a useful contribution to O-RAN slicing: it targets percentile/probabilistic delay guarantees rather than average delay, identifies a concrete training pathology in Lagrangian reward shaping (the 'unlearnable region'), and introduces a performance-based personalization mechanism that is a reasonable alternative to FedAV when client environments are heterogeneous. The comparison against fixed, heuristic, and average-delay baselines is appropriate, and the hardware-in-the-loop validation is a strength. However, the central empirical claims are currently supported by a single simulation run and by a shaped reward whose coefficients are never disclosed, which limits reproducibility and generality.

major comments (5)
  1. [Section 6, Eq. (11)] The coefficients zeta_p, nu_p, gamma_p, zeta_n, nu_n, gamma_n and the clipping bound R_max are never given, and no tuning procedure is described. Because Eq. (11) is the training objective for PDA-DRL, the results in Table 1 cannot be independently reproduced, and the sensitivity of the reported 1% QoS margin, 38% average-delay reduction, and 33% delay-STD reduction to these coefficients is unknown. Please report the values used and, ideally, a parameter-sensitivity study.
  2. [Section 6, Fig. 3(a)] The claim that the practical reward in Eq. (11) preserves the global optimum of the Lagrangian reward in Eq. (22) is supported only by a single scenario (Dmax = 5 ms, epsilon = 0.1). The shaping in Eq. (11) is nonlinear and can in principle shift the argmax in other environments. The paper should either prove conditions under which the optimum is preserved or verify equality of the optimizers across the full set of simulated environments, including the 10 environments used in Fig. 6.
  3. [Section 5.2, Eq. (9)] The action-space definition is internally inconsistent. The set {-2^J, ..., -1, 0, 1, ..., 2^J}, read as consecutive integers, contains 2^{J+1}+1 elements, while the power-of-two reading contains 2J+3 elements; the paper states |A| = 2J+1. Since Section 8 sets J = 5, the number of actions and therefore the output-layer dimension of the policy network are ambiguous. Please correct the formula and report the exact action set used in both the simulation and the testbed experiments.
  4. [Section 8, Table 1] All headline results are point estimates from a single training run. No seeds, confidence intervals, or significance tests are reported, so the claimed 38% and 33% reductions and the 1% QoS margin cannot be separated from training variance. The paper should report multi-seed statistics (e.g., mean and standard deviation over at least 5 seeds) for all policies in Table 1 and Fig. 6.
  5. [Appendix C vs Section 8] The Colosseum validation uses Deep Q-learning with a different action set ({-9, -6, -3, 0, 3, 6, 9}), a different control period (250 ms), and a different exploration schedule, whereas the main simulation method is a policy-gradient agent using the action space of Eq. (9). The appendix is therefore a related feasibility study rather than a validation of the specific PDA-DRL algorithm. Please clarify the relationship and, if the appendix is intended as a validation of PDA-DRL, reconcile the algorithm and hyperparameter choices.
minor comments (5)
  1. [Figure 4] The caption and legend use generic labels 'DRL', 'Fixed', and 'Heuristic', while the text refers to PDA-DRL, MD-DRL, Fixed-Av, and Fixed-Max. Please make the labels consistent.
  2. [Eq. (11) and surrounding text] The text says that when Delta < 0 the first term contributes linearly to the reward, but the expression is Delta^{gamma_n}; please specify whether gamma_n = 1 (and analogously gamma_p = 1) or explain the discrepancy.
  3. [Section 7.3, Eq. (16)] The evaluation procedure for obtaining R^T_i,j is under-specified: it is not stated whether the tested policies are deterministic, how many evaluation trajectories are averaged, or how the positive bias added to the (negative) rewards in Fig. 6 is chosen. This makes the personalization comparison difficult to reproduce.
  4. [Section 5.2 and Section 8] The action-space example A = {10, 20, ..., 120} is inconsistent with the later statement that the RAN has 150 total PRBs. Please clarify whether the actions are in units of PRBs or resource block groups and whether the action range depends on the total PRB budget.
  5. [Section 3 and Section 8] The duration Tb is defined as a PRB time slot in Section 3 and set to 1 ms in Section 8, but Appendix C states that a PRB has a duration of 0.5 ms. Please use consistent time-scale definitions across the simulation and the testbed sections.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reward and personalization claims rest on a standard Lagrangian-reward construction and external empirical comparisons; missing shaping-coefficient values are a reproducibility concern, not a circular reduction.

full rationale

The reward derivation chain is self-contained and not circular. In Appendix A, the per-packet reward is constructed from delay-deadline satisfaction, and the LLN approximation leads to Eq. (21), which is algebraically identical to the Lagrangian dual form in Eq. (22) after choosing u1 = lambda*epsilon and u0 = -lambda*(1-epsilon). This is a standard equivalent construction, not a prediction derived from its own conclusion. The practical reward in Eq. (11) is explicitly a shaped surrogate with free coefficients zeta, nu, gamma, and clipping bound R_max; the paper does not disclose these values and demonstrates argmax preservation in only one scenario (Fig. 3a). That is a legitimate reproducibility and generalization weakness, but it is not circularity: no equation forces the shaped reward to equal the Lagrangian reward, and the reported Table 1 gains are empirical outcomes from training, not quantities implied by the reward definition alone. The personalization method in Eq. (16) selects aggregation weights by testing each candidate model on the target environment and then reports performance on that same environment; this by-design evaluation has an oracle-like flavor, but the comparison against federated averaging and similarity-based aggregation is external and does not reduce a claimed prediction to a fitted input. Self-citations, such as Tehrani et al. 2021 for federated averaging, are used as baselines and comparisons, not as load-bearing justification for the central claims. Overall, the derivation chain is not circular; the main concerns are missing parameter values, single-scenario verification of the shaping assumption, and lack of multi-seed statistics, which are correctness and reproducibility issues rather than circularity.

Assumptions & free parameters 8 free parameters · 5 assumptions · 0 invented entities

The central claims rest on a set of undisclosed reward-shaping constants and a single-case validation that the shaped reward preserves the original optimum. The LLN and Markovian assumptions are standard but unverified in the target environment.

free parameters (8)
  • lambda (Lagrangian multiplier) = not reported
    Trade-off between PRB usage and QoS violation in reward Eq. 10/22; no value or tuning schedule given.
  • zeta_p, nu_p, zeta_n, nu_n = not reported
    Exponential shaping coefficients in practical reward Eq. 11; authors state they 'can be set' but give no values or method.
  • gamma_p, gamma_n = not reported
    Exponents on Delta in practical reward Eq. 11; not specified.
  • R_max = not reported
    Clipping bound for reward in Section 6; value not given.
  • beta = 3
    Temperature in softmax personalization Eq. 16 (Section 8).
  • T = 10
    Number of evaluation episodes for reward-based personalization (Section 8).
  • J = 5
    Action exponent in Eq. 9, though Eq. 9's cardinality formula is inconsistent with the set definition.
  • sigma_m = not reported
    Temperature parameter in feature distance Eq. 13; said to be 'proportional to standard deviation' but not defined.
assumptions (5)
  • domain assumption LLN approximation: the ratio E[|Z|]/E[n_a] is approximated by Pr(d_q < D_max|s) for Ts >> Tb
    Used in Appendix A to derive Eq. 21; requires near-independence and long time slots.
  • ad hoc to paper The modified reward function in Eq. 11 has the same global optimum as the Lagrangian reward Eq. 22 in every environment
    Only demonstrated for a single scenario in Fig. 3a; not proven analytically.
  • standard math The MDP has a steady-state distribution p_{pi_theta}(s) (Eq. 19) and the environment is Markovian in the chosen state features
    Standard DRL assumption; state features in Section 5.1 are assumed sufficient.
  • domain assumption The lower-level scheduler can deliver bits as described in Eq. (3) whenever enough PRBs are allocated
    The scheduling algorithm is not modeled or validated beyond the described bit-rate table.
  • domain assumption Agents can evaluate other agents' models on their own environment for T episodes without impacting the network
    Required by reward-based personalization Section 7.3; assumes testbed allows this and rewards are stationary.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Percentile-Based Deep Reinforcement Learning and Reward Based Personalization For Delay Aware RAN Slicing in O-RAN." pith.science (2026). https://pith.science/paper/VMUA6A5I

@misc{pith2026250718111,
  author       = {Pith},
  title        = {Pith review of: Percentile-Based Deep Reinforcement Learning and Reward Based Personalization For Delay Aware RAN Slicing in O-RAN},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VMUA6A5I}},
  note         = {Machine review of arXiv:2507.18111}
}
read the original abstract

In this paper, we tackle the challenge of radio access network (RAN) slicing within an open RAN (O-RAN) architecture. Our focus centers on a network that includes multiple mobile virtual network operators (MVNOs) competing for physical resource blocks (PRBs) with the goal of meeting probabilistic delay upper bound constraints for their clients while minimizing PRB utilization. Initially, we derive a reward function based on the law of large numbers (LLN), then implement practical modifications to adapt it for real-world experimental scenarios. We then propose our solution, the Percentile-based Delay-Aware Deep Reinforcement Learning (PDA-DRL), which demonstrates its superiority over several baselines, including DRL models optimized for average delay constraints, by achieving a 38\% reduction in resultant average delay. Furthermore, we delve into the issue of model weight sharing among multiple MVNOs to develop a robust personalized model. We introduce a reward-based personalization method where each agent prioritizes other agents' model weights based on their performance. This technique surpasses traditional aggregation methods, such as federated averaging, and strategies reliant on traffic patterns and model weight distance similarities.

Figures

Figures reproduced from arXiv: 2507.18111 by the authors.

Figure 1
Figure 1. A network containing a central controller (telco [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. RL scheme for the RAN slicing problem. 4. Problem Formulation The downlink SNR of user j in cell k at time slot t can be defined as: γ t j,k = P t k g t j,k Nj (1) where g t j,k is the channel gain between the kth RAN and jth user at time slot t, Pt is the kth RAN’s RF transmit power and Nj is the noise power at the user j terminal. It is assumed that the each RAN is using uniform power allocation for all PRBs. We a… view at source ↗
Figure 3
Figure 3. (a) Theory vs practical reward functions. (b) [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: Performance comparison of DRL-based methods, [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Convergence plots of different features in two dis [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Comparison between different aggregation methods. positive bias to the obtained rewards, as they are inherently negative. For the aggregation coefficient in (16), we set β = 3 and T = 10. In Appendix C, we have validated our DRL solution through experimental framework …
Figure 7
Figure 7. Figure 7: ORAN compliant system architecture. B. ORAN Compliance In [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]
Figure 8
Figure 8. Figure 8: Demonstration of signal transmission in Colosseu [PITH_FULL_IMAGE:figures/full_fig_p015_8.png]
Figure 9
Figure 9. Figure 9: Experimental framework architecture. 16 [PITH_FULL_IMAGE:figures/full_fig_p016_9.png]
Figure 10
Figure 10. Figure 10: Convergence plots in experimental setting. [PITH_FULL_IMAGE:figures/full_fig_p017_10.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 42 canonical work pages

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Federated deep reinforcement learning for open ran slicing in 6g networks

    Abouaomar, A., Taik, A., Filali, A., and Cherkaoui, S. Federated deep reinforcement learning for open ran slicing in 6g networks. IEEE Communications Magazine, 2022

    work page 2022

  3. [3]

    Energy saving and traffic steering use case and testing by o-ran ric xapp/rapp multi-vendor interoperability

    Akman, A., Oliver, P., Jones, M., Tehrani, P., Hoffmann, M., and Li, J. Energy saving and traffic steering use case and testing by o-ran ric xapp/rapp multi-vendor interoperability. In 2024 IEEE 100th Vehicular Technology Conference (VTC2024-Fall), pp.\ 1--6. IEEE, 2024

    work page 2024

  4. [4]

    A., Ksentini, A., and Bouaziz, M

    Bakri, S., Frangoudis, P. A., Ksentini, A., and Bouaziz, M. Data-driven ran slicing mechanisms for 5g and beyond. IEEE Transactions on Network and Service Management, 18 0 (4): 0 4654--4668, 2021

    work page 2021

  5. [5]

    Scope: An open and softwarized prototyping platform for nextg systems

    Bonati, L., D'Oro, S., Basagni, S., and Melodia, T. Scope: An open and softwarized prototyping platform for nextg systems. In Proceedings of the 19th Annual International Conference on Mobile Systems, Applications, and Services, pp.\ 415--426, 2021 a

    work page 2021

  6. [6]

    Intelligence and learning in o-ran for data-driven nextg cellular networks

    Bonati, L., D'Oro, S., Polese, M., Basagni, S., and Melodia, T. Intelligence and learning in o-ran for data-driven nextg cellular networks. IEEE Communications Magazine, 59 0 (10): 0 21--27, 2021 b

    work page 2021

  7. [7]

    Colosseum: Large-scale wireless experimentation through hardware-in-the-loop network emulation

    Bonati, L., Johari, P., Polese, M., D’Oro, S., Mohanti, S., Tehrani-Moayyed, M., Villa, D., Shrivastava, S., Tassie, C., Yoder, K., et al. Colosseum: Large-scale wireless experimentation through hardware-in-the-loop network emulation. In 2021 IEEE International Symposium on Dynamic Spectrum Access Networks (DySPAN), pp.\ 105--113. IEEE, 2021 c

    work page 2021

  8. [8]

    User access control in open radio access networks: A federated deep reinforcement learning approach

    Cao, Y., Lien, S.-Y., Liang, Y.-C., Chen, K.-C., and Shen, X. User access control in open radio access networks: A federated deep reinforcement learning approach. IEEE Transactions on Wireless Communications, 21 0 (6): 0 3721--3736, 2021

    work page 2021

Show all 44 references
  1. [9]

    Orchestran: Network automation through orchestrated intelligence in the open ran

    D’Oro, S., Bonati, L., Polese, M., and Melodia, T. Orchestran: Network automation through orchestrated intelligence in the open ran. In IEEE INFOCOM 2022-IEEE Conference on Computer Communications, pp.\ 270--279. IEEE, 2022

    2022

  2. [10]

    E., Jemaa, S

    Elayoubi, S. E., Jemaa, S. B., Altman, Z., and Galindo-Serrano, A. 5g ran slicing for verticals: Enablers and challenges. IEEE Communications Magazine, 57 0 (1): 0 28--34, 2019

    2019

  3. [11]

    Radio access network (ran) digital twins

    Ericsson. Radio access network (ran) digital twins. White paper, 2023

    2023

  4. [12]

    Dynamic sdn-based radio access network slicing with deep reinforcement learning for urllc and embb services

    Filali, A., Mlika, Z., Cherkaoui, S., and Kobbane, A. Dynamic sdn-based radio access network slicing with deep reinforcement learning for urllc and embb services. IEEE Transactions on Network Science and Engineering, 9 0 (4): 0 2174--2187, 2022

    2022

  5. [13]

    Learning based on graph: A joint interference coordination for cluster-wise distributed mu-mimo

    Ge, C., Xia, S., Chen, Q., and Adachi, F. Learning based on graph: A joint interference coordination for cluster-wise distributed mu-mimo. IEEE Communications Letters, 2023

    2023

  6. [14]

    Gan-powered deep distributional reinforcement learning for resource management in network slicing

    Hua, Y., Li, R., Zhao, Z., Chen, X., and Zhang, H. Gan-powered deep distributional reinforcement learning for resource management in network slicing. IEEE Journal on Selected Areas in Communications, 38 0 (2): 0 334--349, 2019

    2019

  7. [15]

    Personalized cross-silo federated learning on non-iid data

    Huang, Y., Chu, L., Zhou, Z., Wang, L., Liu, J., Pei, J., and Zhang, Y. Personalized cross-silo federated learning on non-iid data. In Proceedings of the AAAI Conference on Artificial Intelligence, 2021

    2021

  8. [16]

    E., and Erol-Kantarci, M

    Joda, R., Pamuklu, T., Iturria-Rivera, P. E., and Erol-Kantarci, M. Deep reinforcement learning-based joint user association and cu-du placement in o-ran. IEEE Transactions on Network and Service Management, 2022

    2022

  9. [17]

    and Iosifidis, G

    Kalntis, M. and Iosifidis, G. Energy-aware scheduling of virtualized base stations in o-ran with online learning. In GLOBECOM 2022-2022 IEEE Global Communications Conference, pp.\ 6048--6054. IEEE, 2022

    2022

  10. [18]

    Generative ai in mobile networks: a survey

    Karapantelakis, A., Alizadeh, P., Alabassi, A., Dey, K., and Nikou, A. Generative ai in mobile networks: a survey. Annals of Telecommunications, pp.\ 1--19, 2023

    2023

  11. [19]

    K., Lazaridis, P

    Kougioumtzidis, G., Vlahov, A., Poulkov, V. K., Lazaridis, P. I., and Zaharis, Z. D. Deep learning-aided qoe prediction for virtual reality applications over open radio access networks. IEEE Access, 2023

    2023

  12. [20]

    S., Singh, T., Zugno, T., Cuomo, F., and Melodia, T

    Lacava, A., Polese, M., Sivaraj, R., Soundrarajan, R., Bhati, B. S., Singh, T., Zugno, T., Cuomo, F., and Melodia, T. Programmable and customized intelligence for traffic steering in 5g networks using open ran architectures. arXiv preprint arXiv:2209.14171, 2022

    2022 arXiv

  13. [21]

    Session management for urllc in 5g open radio access network: A machine learning approach

    Lien, S.-Y., Deng, D.-J., and Chang, B.-C. Session management for urllc in 5g open radio access network: A machine learning approach. In 2021 International Wireless Communications and Mobile Computing (IWCMC), pp.\ 2050--2055. IEEE, 2021

    2021

  14. [22]

    McMahan, B., Moore, E., Ramage, D., Hampson, S., and y Arcas, B. A. Communication-efficient learning of deep networks from decentralized data. In Artificial intelligence and statistics, pp.\ 1273--1282. PMLR, 2017

    2017

  15. [23]

    B., and Abou-Zeid, H

    Mei, J., Wang, X., Zheng, K., Boudreau, G., Sediq, A. B., and Abou-Zeid, H. Intelligent radio access network slicing for service provisioning in 6g: A hierarchical deep reinforcement learning approach. IEEE Transactions on Communications, 69 0 (9): 0 6063--6078, 2021

    2021

  16. [24]

    B., Quang, P

    Messaoud, S., Bradai, A., Ahmed, O. B., Quang, P. T. A., Atri, M., and Hossain, M. S. Deep federated q-learning-based network slicing for industrial iot. IEEE Transactions on Industrial Informatics, 17 0 (8): 0 5572--5582, 2020

    2020

  17. [25]

    A., Veness, J., Bellemare, M

    Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A. A., Veness, J., Bellemare, M. G., Graves, A., Riedmiller, M., Fidjeland, A. K., Ostrovski, G., et al. Human-level control through deep reinforcement learning. nature, 518 0 (7540): 0 529--533, 2015

    2015

  18. [26]

    K., Shah-Mansouri, V., Parsaeefard, S., and L \'o pez, O

    Motalleb, M. K., Shah-Mansouri, V., Parsaeefard, S., and L \'o pez, O. L. A. Resource allocation in an open ran system using network slicing. IEEE Transactions on Network and Service Management, 20 0 (1): 0 471--485, 2022

    2022

  19. [27]

    M., Abou-zeid, H., and Hassanein, H

    Nagib, A. M., Abou-zeid, H., and Hassanein, H. S. Accelerating reinforcement learning via predictive policy transfer in 6g ran slicing. IEEE Transactions on Network and Service Management, 2023

    2023

  20. [28]

    S., Singh, S., Banerji, R., Reed, J

    Niknam, S., Roy, A., Dhillon, H. S., Singh, S., Banerji, R., Reed, J. H., Saxena, N., and Yoon, S. Intelligent o-ran for beyond 5g and 6g wireless networks. In 2022 IEEE Globecom Workshops (GC Wkshps), pp.\ 215--220. IEEE, 2022

    2022

  21. [29]

    O-RAN Architecture Description 10.0

    O-RAN Alliance . O-RAN Architecture Description 10.0 . Technical Specification (TS) O-RAN.WG1.OAD-R003-v10.00, O-RAN Alliance, Oct 2023. R003. [Online]. Available: https://www.o-ran.org/specifications

    2023

  22. [30]

    N., Tetzlaff, T., Nassar, M., Nikopour, H., and Talwar, S

    Orhan, O., Swamy, V. N., Tetzlaff, T., Nassar, M., Nikopour, H., and Talwar, S. Connection management xapp for o-ran ric: A graph neural network and reinforcement learning approach. In 2021 20th IEEE International Conference on Machine Learning and Applications (ICMLA), pp.\ 9...

    2021

  23. [31]

    Understanding o-ran: Architecture, interfaces, algorithms, security, and research challenges

    Polese, M., Bonati, L., D’Oro, S., Basagni, S., and Melodia, T. Understanding o-ran: Architecture, interfaces, algorithms, security, and research challenges. IEEE Communications Surveys & Tutorials, 2023

    2023

  24. [32]

    F., Simeone, O., and Durisi, G

    Popovski, P., Trillingsgaard, K. F., Simeone, O., and Durisi, G. 5g wireless network slicing for embb, urllc, and mmtc: A communication-theoretic view. Ieee Access, 6: 0 55765--55779, 2018

    2018

  25. [33]

    Queue-learning: A reinforcement learning approach for providing quality of service

    Raeis, M., Tizghadam, A., and Leon-Garcia, A. Queue-learning: A reinforcement learning approach for providing quality of service. arXiv preprint arXiv:2101.04627, 2021

    2021 arXiv

  26. [34]

    S., Demir, U., Stephenson, N., Soltani, N., Shah, V

    Reus-Muns, G., Upadhyaya, P. S., Demir, U., Stephenson, N., Soltani, N., Shah, V. K., and Chowdhury, K. R. Senseoran: O-ran based radar detection in the cbrs band. IEEE Journal on Selected Areas in Communications, 2023

    2023

  27. [35]

    On the specialization of fdrl agents for scalable and distributed 6g ran slicing orchestration

    Rezazadeh, F., Zanzi, L., Devoti, F., Chergui, H., Costa-P \'e rez, X., and Verikoukis, C. On the specialization of fdrl agents for scalable and distributed 6g ran slicing orchestration. IEEE Transactions on Vehicular Technology, 2022

    2022

  28. [36]

    Setayesh, M., Bahrami, S., and Wong, V. W. Resource slicing for embb and urllc services in radio access network using hierarchical deep learning. IEEE Transactions on Wireless Communications, 21 0 (11): 0 8950--8966, 2022

    2022

  29. [37]

    K., and Radunovic, B

    Sun, C., Pawar, U., Khoja, M., Foukas, X., Marina, M. K., and Radunovic, B. Spotlight: Accurate, explainable and efficient anomaly detection for open ran. In Proceedings of the 30th Annual International Conference on Mobile Computing and Networking, pp.\ 923--937, 2024 a

    2024

  30. [38]

    Research report on digital twin ran use cases, May 2024 b

    Sun, K., Tao, C., Johari, P., D’Oro, S., Polese, M., Bonati, L., Melodia, T., Rajendran, G., Kundu, N., Chakfeh, Y., Raghothaman, B., D'angelo, M., Agarwal, R., Kundu, L., Lin, X., and Dick, C. Research report on digital twin ran use cases, May 2024 b . Report ID: RR-2024-07

    2024

  31. [39]

    Federated deep reinforcement learning for the distributed control of nextg wireless networks

    Tehrani, P., Restuccia, F., and Levorato, M. Federated deep reinforcement learning for the distributed control of nextg wireless networks. In 2021 IEEE International Symposium on Dynamic Spectrum Access Networks (DySPAN), pp.\ 248--253. IEEE, 2021

    2021

  32. [40]

    Off-policy learning in contextual bandits for remote electrical tilt optimization

    Vannella, F., Jeong, J., and Proutiere, A. Off-policy learning in contextual bandits for remote electrical tilt optimization. IEEE Transactions on Vehicular Technology, 72 0 (1): 0 546--556, 2022

    2022

  33. [41]

    and Liyanage, M

    Wijethilaka, S. and Liyanage, M. Survey on network slicing for internet of things realization in 5g networks. IEEE Communications Surveys & Tutorials, 23 0 (2): 0 957--994, 2021

    2021

  34. [42]

    Dynamic ran slicing for service-oriented vehicular networks via constrained learning

    Wu, W., Chen, N., Zhou, C., Li, M., Shen, X., Zhuang, W., and Li, X. Dynamic ran slicing for service-oriented vehicular networks via constrained learning. IEEE Journal on Selected Areas in Communications, 39 0 (7): 0 2076--2089, 2020

    2020

  35. [43]

    M., Dzaferagic, M., Collins, D., Comarela, G., Martinello, M., and Ruffini, M

    Xavier, B. M., Dzaferagic, M., Collins, D., Comarela, G., Martinello, M., and Ruffini, M. Machine learning-based early attack detection using open ran intelligent controller. arXiv preprint arXiv:2302.01864, 2023

    2023 arXiv

  36. [44]

    Q., Chen, J., Cao, X., and Wu, D

    Yang, P., Xi, X., Quek, T. Q., Chen, J., Cao, X., and Wu, D. Ran slicing for massive iot and bursty urllc service multiplexing: Analysis and optimization. IEEE Internet of Things Journal, 8 0 (18): 0 14258--14275, 2021

    2021

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.