DORA: Dynamic O-RAN Resource Allocation for Multi-Slice 5G Networks

Alireza Ebrahimi Dorcheh; Fatemeh Afghah; Tolunay Seyfi

REVIEW 4 major objections 3 minor 1 cited by

DORA: Dynamic O-RAN Resource Allocation for Multi-Slice 5G Networks

T0 review · 4 major / 3 minor · reviewed 2026-08-04 · deepseek-v4-flash

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

Pith's one-line read A single PPO agent can split 5G radio blocks across three slices live

desk verdict DORA is a genuine OAI engineering effort, but its headline 'first fully online' and 'outperforms on all KPIs' claims collapse under its own results. read the letter →

arxiv 2509.07242 v1 pith:DWXNKUSO submitted 2025-09-08 cs.NI

Alireza Ebrahimi Dorcheh , Tolunay Seyfi , Fatemeh Afghah This is my paper · ORCID

classification cs.NI

keywords O-RANnetworkslicingphysicalresourceblockallocationdeepreinforcementlearningPPOOpenAirInterface5Gmanagementservice-levelagreements

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

Network operators slice one 5G cell into three service classes—ultra-reliable low-latency, mobile broadband, and massive IoT—that compete for the same 106 physical resource blocks (the smallest schedulable radio resource) every transmission interval. The paper claims that a single PPO-based reinforcement-learning agent, trained offline on an OAI-derived throughput table and evaluated online in the OpenAirInterface stack, allocates these blocks better than three fixed or heuristic baselines and a DQN agent: lower URLLC latency, higher eMBB throughput with fewer SLA violations, and wider mMTC coverage, all under deliberate congestion. The key design choice is decoupling slice-level allocation from per-user scheduling—the RL agent decides slice totals, then a deterministic round-robin assigns blocks inside each slice. If the claim holds, DORA is a deployable O-RAN xApp that lets operators adapt slicing in real time without sacrificing any one service class.

What carries the argument

Key machinery: (1) a discrete-control MDP whose action space is every integer triple (a,b,c) with a+b+c=106 PRBs—3,003 actions; (2) a PPO policy trained offline against a throughput lookup table indexed by (pathloss, PRBs) so training avoids expensive PHY simulation; (3) a weighted reward that maps SLA violations into bounded costs; (4) deterministic round-robin inside each slice, so the RL agent only decides slice totals, not user-level schedules; (5) an E2-like Python bridge to the OAI stack for closed-loop telemetry and control.

What would settle it

Re-run DORA's online evaluation while replacing the lookup table with live OAI throughput measurements, or add fading and measurement noise during pretraining. If the reported URLLC latency, eMBB throughput, and mMTC coverage margins over the baselines collapse or reverse, the transfer-from-table assumption is the load-bearing failure.

Watch

Extended reading notes

Core claim

DORA's central claim: one PPO agent can perform adaptive slice-level PRB allocation entirely online in O-RAN, without jointly scheduling individual users. It observes slice type, normalized demand, and pathloss for each URLLC/eMBB user, picks one of 3,003 integer PRB triples summing to 106, and earns a weighted reward (0.5 URLLC latency, 0.4 eMBB throughput, 0.1 mMTC coverage). Offline pretraining uses an OAI-derived throughput lookup table; online evaluation closes the loop in the OAI simulator, adding fading, noise, and contention. The paper reports DORA beats three non-learning baselines and DQN under congestion while preserving URLLC priority.

Load-bearing premise

The whole evaluation assumes that the offline lookup table, built from controlled OAI runs, captures the throughput the online OAI environment will actually deliver; the paper lists fading, measurement noise, and cross-slice contention as differences but never measures the gap.

Editorial extensions

If this is right

If DORA is right, operators can run slice-level PRB allocation as a closed-loop xApp on a near-real-time RIC without dedicated PHY hardware, using software UEs for testing.
Decoupling slice-level decisions from per-UE scheduling means the agent design can scale to more UEs inside a slice: the RL action space stays at slice granularity.
Under overload, the learned policy will occasionally miss the URLLC latency target by a small margin to free resources for eMBB and mMTC—a deliberate trade-off that static allocations cannot encode.
The reported results make DORA a stronger online alternative to offline-trained DRL xApps, since it is pretrained quickly on a table and then evaluated against live OAI dynamics.

Reading between the lines

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

The transfer from lookup-table pretraining to live OAI is the paper's unmeasured risk; a natural next experiment is to compare table-expected vs OAI-measured throughput per (pathloss, PRB) pair and to test online fine-tuning.
Because the action space enumerates every integer split of 106 PRBs, the same MDP could be re-run with different reward weights to trace the operator's Pareto frontier among the three slices; DORA reports only one weighting (0.5/0.4/0.1).
The E2-like Python bridge with memory-mapped files is a simplification of the O-RAN E2AP standard; production deployment would require replacing that bridge with a full E2 interface, which the paper does not implement.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Two things to know before you read this one. The abstract claims DORA is the first fully online DRL framework for slice-aware PRB allocation in O-RAN and that it outperforms three baselines plus DQN on all three slice KPIs. Neither claim survives contact with the paper's own text. Section II describes training as offline with online evaluation only, and the cited REAL system is an online PPO xApp. Section VII reports that Hard Slicing has the best URLLC SLA compliance, DQN has a slight edge in eMBB throughput, and Priority-based/DQN lead mMTC coverage. DORA's actual contribution is 'most balanced,' not 'best at everything.'

What is genuinely here: a complete, careful PPO-based slice-level PRB allocator paired with round-robin intra-slice scheduling, pretrained on an OAI-derived throughput lookup table and evaluated closed-loop in the OAI RFsim stack. That pipeline is a plausible, reusable template for O-RAN xApp development. The decoupling of slice-level control from per-UE scheduling is sensible and well motivated. The evaluation is detailed, with CDFs for latency and throughput and explicit discussion of trade-offs. The authors also acknowledge the hard-slicing trade-off and the priority baseline's mismatch, which suggests they understand their own results.

The soft spots are real but mostly claim-level rather than implementation-level. The reward weights (0.5, 0.4, 0.1) are hand-picked, and the 'intelligent balance' is largely a consequence of those weights, not emergent learning. mMTC coverage is defined as floor(PRB/5), so the metric is a direct arithmetic consequence of allocation. There are no error bars or multiple seeds, no released code or data, and only one congested regime against weak baselines. The biggest issue is that the paper's selling points are contradicted by its own evaluation. That is not a minor copyediting problem; the framing needs to change substantially.

Who should read it: anyone building an OAI-based slice scheduler will find the integration details useful. It is a solid engineering report, but as a novelty claim about being first, it doesn't hold up.

Recommendation: The empirical work is real enough that I would not desk-reject it; I'd send it to review with a note to the referee to check the claims against Section VII. The authors should be pushed to revise the abstract and introduction to describe DORA as a balanced trade-off engine with a reproducible OAI implementation, not as a universal winner.

Referee Report

4 major / 3 minor

Summary. The paper proposes DORA, a PPO-based DRL framework for slice-level PRB allocation in O-RAN, targeting URLLC, eMBB, and mMTC slices. The agent allocates PRBs based on observed traffic and channel conditions, while intra-slice scheduling is deterministic round-robin. Training is performed offline with an OAI-derived throughput lookup table, followed by online evaluation in OpenAirInterface. The paper compares DORA with three non-learning baselines (Hard Slicing, Priority-based, Fair Active-User) and a DQN agent, reporting URLLC latency, eMBB throughput, and mMTC coverage. The headline claims are that DORA is the first fully online DRL framework for adaptive slice-aware PRB allocation and that it outperforms all baselines on all three slice KPIs.

Significance. If the claims were fully supported, DORA would be a useful contribution: an OAI-integrated, slice-level DRL xApp with a reduced-complexity scheduling interface and a comparison against multiple baselines. The paper's strengths include the use of a realistic OAI stack, explicit treatment of a 14-UE scenario, and distributional KPI reporting (CDFs). However, as detailed below, the paper's own results contradict the central superiority claim: DORA is not best on any single slice KPI, but rather a balanced compromise. In addition, the 'fully online' claim is internally inconsistent with the offline-training description. The significance of the contribution is therefore substantially weaker than the abstract and introduction suggest, though the underlying framework and experimental setup could support a revised, less overreaching narrative.

major comments (4)

[Abstract and §VII] The central claim that DORA 'achieves lower URLLC latency, higher eMBB throughput with fewer SLA violations, and broader mMTC coverage' is not supported by the results. Section VII-A states that at the 400 ms SLA threshold 'the Hard Slicing baseline exhibits the highest probability,' with DORA following closely. Section VII-B says the DRL agents are 'very similar, with DQN holding a slight edge.' Section VII-C says 'the Priority-based and DQN agents achieve the best performance' for mMTC, while DORA provides only 'comparable and effective coverage.' Thus DORA is best on no individual KPI; it is the most balanced. The headline must be revised to the supported claim, and the authors should provide a quantitative multi-objective comparison (e.g., a single utility metric) if they wish to claim superiority in trade-off handling.
[Abstract, §I, §II, §IV-A] The 'fully online' characterization is internally inconsistent. The abstract says DORA 'supports online training and adapts continuously,' and the introduction claims 'the first fully online DRL framework.' However, Section II states that 'training is performed offline,' and Section IV-A describes a hybrid pipeline with 'Offline Training' followed by 'Online Evaluation.' No online policy updates or fine-tuning are performed. This is a load-bearing inconsistency because the stated novelty depends on online training. Please clarify whether the contribution is online evaluation of an offline-trained policy, and if online training is actually supported, provide the mechanism and evidence.
[§VI-C] The offline-to-online transfer of the pretrained policy is a key unvalidated assumption. Section VI-C lists stochastic fading, measurement noise, and cross-slice contention as differences 'absent from offline approximations,' but the paper provides no evaluation of how representative the offline lookup table is, no comparison of offline vs. online reward or KPI distributions, and no online adaptation. Consequently, the reported online KPIs do not isolate the quality of the learned policy from possible lookup-table inaccuracies. An ablation (e.g., training with perturbed tables, or measuring the offline-online performance gap) is needed before the online results can be interpreted as validating the DORA agent.
[Eq. (3)–(6), §IV-C, §VII-D] The reward weights (w_U, w_E, w_M) = (0.5, 0.4, 0.1) are fixed, and Section VII-D attributes DORA's balanced behavior to these weights. The paper presents no sensitivity analysis, so the conclusion that DORA 'intelligently balances' the objectives is partly forced by construction. Also, Eq. (6) with N_serv = floor(PRB_mMTC/5) makes the mMTC reward a deterministic affine function of the allocated PRBs; 'coverage' is therefore just a relabeled PRB count, not a device-level QoS measure. These limitations should be stated, and the interpretation of Figure 4 should be adjusted accordingly.

minor comments (3)

[Throughout] There are several typos and spacing errors, e.g., 'aDQNagent' in the abstract and 'aDQN agent' in Section VII. Please proofread the text.
[§IV-A] The paper claims 'standards-compliant' integration but then says DORA uses a Python-based 'E2-like' control architecture with memory-mapped files, not full E2AP. This should be clarified: is the implementation E2-compliant or only E2-inspired? The wording in the abstract and introduction may overstate compliance.
[§I and Table I] The paper states DORA supports 'fully online evaluation with up to four software-based UEs running in parallel' while also describing a 14-UE scenario. Table I's 'Online' column is ambiguous. Please specify what 'online' means in each context (training vs. evaluation vs. UE concurrency).

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the RL pipeline is self-contained, though the paper has internal consistency issues.

full rationale

The paper's central derivation chain is self-contained rather than circular. The PPO agent is trained offline on an OAI-derived lookup table and evaluated online in OAI; the reported URLLC latency, eMBB throughput, and mMTC coverage are measured from the actual closed-loop interaction, not computed from the training table or reward function. The mMTC metric Nserv = floor(PRB_mMTC/5) is a deterministic consequence of the chosen action, and the reward includes the same quantity, so the agent optimizes the KPI it is later scored on; however, this is standard RL objective alignment, not a fitted parameter repackaged as a prediction. No self-citation is load-bearing: REAL [8] is cited for comparison, but DORA's claim of distinction rests on its use of OAI RFsim rather than GNU Radio emulation. The paper does contain serious internal inconsistencies—the abstract claims 'online training' while Section II and Section IV state 'training is performed offline' and only online evaluation occurs; and the abstract's claim of outperforming all baselines on all three KPIs is directly contradicted by Sections VII-A through VII-C (Hard Slicing has the highest URLLC SLA compliance, DQN has a slight edge in eMBB, and Priority-based/DQN achieve the best mMTC coverage). These are factual/consistency defects, not circular reasoning. No derivation step reduces to its own input by definition.

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

The central claim rests on four hand-set constants and two unvalidated transfer assumptions: lookup-table fidelity and offline-to-online transfer. No invented physical entities are introduced. The ledger is small, but the constants directly shape the headline trade-off result, especially the reward weights and the mMTC PRB demand formula.

free parameters (4)

Reward weights (w_U, w_E, w_M) = (0.5, 0.4, 0.1)
Hand-chosen in Eq. (3); directly determine the trade-off DORA learns, so the claim that DORA balances slices is partly an artifact of these weights.
Pathloss calibration offset C_off = 83.84 dB
Calibrated to match "physical-layer measurements in our experimental setup" (Eq. 1); the RL observation vector depends on normalized pathloss.
Efficiency factor eta = approximately 0.7
Listed in Table II as a simulation parameter with no derivation or fitting procedure; it scales the throughput model.
mMTC PRB demand per device = 5 PRBs/device
Chosen in Sections III-A and IV-C; it defines N_serv = floor(PRB_mMTC/5), so the mMTC coverage KPI is a deterministic function of the action.

assumptions (4)

domain assumption The OAI-derived throughput lookup table accurately predicts throughput for arbitrary (pathloss, PRB) pairs.
Introduced in Section IV-B; offline pretraining and the priority baseline both rely on it; no validation against online measured values is reported.
ad hoc to paper The reward function (Eqs. 3-6) captures the operator's true multi-slice objective.
Weights and SLA thresholds are chosen by the authors; the paper does not derive them from any operator utility or 3GPP standard.
domain assumption The offline-trained policy transfers to online OAI without requiring online adaptation.
Assumed in Section VI-C, which lists stochastic fading, measurement noise, and cross-slice contention as differences but does not quantify their impact on the trained policy.
domain assumption The pathloss model with calibration offset (Eq. 1) matches OAI physics sufficiently for the observation vector.
Stated in Section III-B without supporting data; pathloss values feed the RL observation and the lookup table.

how reviews work

0 comments

Cite this review

Pith. "Pith review of DORA: Dynamic O-RAN Resource Allocation for Multi-Slice 5G Networks." pith.science (2026). https://pith.science/paper/DWXNKUSO

@misc{pith2026250907242,
  author       = {Pith},
  title        = {Pith review of: DORA: Dynamic O-RAN Resource Allocation for Multi-Slice 5G Networks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DWXNKUSO}},
  note         = {Machine review of arXiv:2509.07242}
}

read the original abstract

The fifth generation (5G) of wireless networks must simultaneously support heterogeneous service categories, including Ultra-Reliable Low-Latency Communications (URLLC), enhanced Mobile Broadband (eMBB), and massive Machine-Type Communications (mMTC), each with distinct Quality of Service (QoS) requirements. Meeting these demands under limited spectrum resources requires adaptive and standards-compliant radio resource management. We present DORA (Dynamic O-RAN Resource Allocation), a deep reinforcement learning (DRL) framework for dynamic slice-level Physical Resource Block (PRB) allocation in Open RAN. DORA employs a PPO-based RL agent to allocate PRBs across URLLC, eMBB, and mMTC slices based on observed traffic demands and channel conditions. Intra-slice PRB scheduling is handled deterministically via round-robin among active UEs, simplifying control complexity and improving training stability. Unlike prior work, DORA supports online training and adapts continuously to evolving traffic patterns and cross-slice contention. Implemented in the standards-compliant OpenAirInterface (OAI) RAN stack and designed for deployment as an O-RAN xApp, DORA integrates seamlessly with RAN Intelligent Controllers (RICs). Extensive evaluation under congested regimes shows that DORA outperforms three non-learning baselines and a \texttt{DQN} agent, achieving lower URLLC latency, higher eMBB throughput with fewer SLA violations, and broader mMTC coverage without starving high-priority slices. To our knowledge, this is the first fully online DRL framework for adaptive, slice-aware PRB allocation in O-RAN.

Figures

Figures reproduced from arXiv: 2509.07242 by the authors.

**Figure 2.** CDF of URLLC end-to-end latency. DORA (PPO) demonstrates strong [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** CDF of the difference between achieved and requested eMBB [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Average number of concurrently serviced mMTC devices. DORA [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

ORAN-DEFEND: Subspace Detection and Sanitization of Backdoor DRL xApps in Open RAN
cs.CR 2026-07 conditional novelty 5.0 of 10

SVD projection onto a clean-KPI subspace recovers 100% DRL return against four O-RAN backdoor attacks whenever trigger energy lies in the orthogonal complement.

Reference graph

Works this paper leans on

13 extracted references · 11 canonical work pages · cited by 1 Pith paper

[1]

Next Generation Cellular Networks Prototyping O-RAN Enabled UA V Exper- imentation for the AERPAW Testbed,

J. Moore, A. S. Abdalla, C. Ueltschey, and V . Marojevic, “Next Generation Cellular Networks Prototyping O-RAN Enabled UA V Exper- imentation for the AERPAW Testbed,”IEEE Communications Magazine, vol. 63, no. 2, pp. 78–84, 2025. doi: 10.1109/MCOM.001.2400327

work page doi:10.1109/mcom.001.2400327 2025
[2]

NexRAN: Closed- loop RAN slicing in POWDER – A top-to-bottom open-source open-RAN use case,

D. Johnson, D. Maas, and J. Van Der Merwe, “NexRAN: Closed- loop RAN slicing in POWDER – A top-to-bottom open-source open-RAN use case,” inProc. WiNTECH, pp. 17–23, 2021. doi: 10.1145/3477086.3480842

work page arXiv 2021
[3]

COSMOS: A city-scale programmable testbed for experimentation with advanced wireless,

D. Raychaudhuriet al., “COSMOS: A city-scale programmable testbed for experimentation with advanced wireless,” inProc. ACM MobiCom, pp. 1–13, 2020

work page 2020
[4]

PandORA: Automated Design and Comprehensive Evaluation of Deep Reinforce- ment Learning Agents for Open RAN,

C. Tsampazi, M. Levorato, F. Restuccia, and T. Melodia, “PandORA: Automated Design and Comprehensive Evaluation of Deep Reinforce- ment Learning Agents for Open RAN,”IEEE Transactions on Mobile Computing, Early Access, 2025. Available: 10766614

work page 2025
[5]

A Comparative Analysis of Deep Reinforcement Learning-Based xApps in O-RAN,

C. Tsampazi, F. Restuccia, and T. Melodia, “A Comparative Analysis of Deep Reinforcement Learning-Based xApps in O-RAN,” inProc. IEEE INFOCOM, 2023. Available: 10437367

work page 2023
[6]

ColO- RAN: Developing Machine Learning-Based xApps for Open RAN Closed-Loop Control on Programmable Experimental Platforms,

M. Polese, L. Bonati, S. D’Oro, S. Sharma, and T. Melodia, “ColO- RAN: Developing Machine Learning-Based xApps for Open RAN Closed-Loop Control on Programmable Experimental Platforms,”IEEE Communications Magazine, 2022. Available: 9814869

work page 2022
[7]

OrchestRAN: Orches- trating Network Intelligence in the Open RAN,

S. D’Oro, M. Polese, L. Bonati, and T. Melodia, “OrchestRAN: Orches- trating Network Intelligence in the Open RAN,”IEEE Communications Magazine, 2024. Available: 10356734

work page 2024
[8]

REAL: Re- inforcement Learning-Enabled xApps for Experimental Closed-Loop Optimization in O-RAN with OSC RIC and srsRAN,

R. Barker, A. E. Dorcheh, T. Seyfi, and F. Afghah, “REAL: Re- inforcement Learning-Enabled xApps for Experimental Closed-Loop Optimization in O-RAN with OSC RIC and srsRAN,”arXiv preprint arXiv:2502.00715, 2025. Available: arXiv:2502.00715

arXiv 2025

Show all 13 references

[9]

Colosseum: The Open RAN Digital Twin,

M. Polese, S. D’Oro, L. Bonati, and T. Melodia, “Colosseum: The Open RAN Digital Twin,”IEEE Communications Magazine, 2024. Available: 10643670

2024
[10]

AdaSlicing: Adaptive Online Network Slicing under Continual Network Dynamics in Open Radio Access Networks,

M. Zhao, Y . Zhang, Q. Liu, A. Kak, and N. Choi, “AdaSlicing: Adaptive Online Network Slicing under Continual Network Dynamics in Open Radio Access Networks,”IEEE INFOCOM, Accepted, 2025. Preprint: arXiv:2501.06943

2025 arXiv
[11]

InSlicing: Inter- pretable Slice Configuration for Open Radio Access Networks,

M. Zhao, Y . Zhang, Q. Liu, A. Kak, and N. Choi, “InSlicing: Inter- pretable Slice Configuration for Open Radio Access Networks,”IEEE INFOCOM, Accepted, 2025. Preprint: arXiv:2502.15918

2025 arXiv
[12]

ORANSlice: An Open Source 5G Network Slicing Platform for O-RAN,

H. Chenget al., “ORANSlice: An Open Source 5G Network Slicing Platform for O-RAN,” inProc. ACM MobiCom, pp. 2297–2302, 2024. doi: 10.1145/3636534.3701544

2024
[13]

OrchestRAN: Or- chestrating Network Intelligence in the Open RAN,

S. D’Oro, L. Bonati, M. Polese, and T. Melodia, “OrchestRAN: Or- chestrating Network Intelligence in the Open RAN,”IEEE Transactions on Mobile Computing, vol. 23, no. 7, pp. 7952–7968, 2024. doi: 10.1109/TMC.2023.3342711

2024

Pith tools

Reviewed August 4, 2026 · model on record in the stance chip above.

[1] [1]

Next Generation Cellular Networks Prototyping O-RAN Enabled UA V Exper- imentation for the AERPAW Testbed,

J. Moore, A. S. Abdalla, C. Ueltschey, and V . Marojevic, “Next Generation Cellular Networks Prototyping O-RAN Enabled UA V Exper- imentation for the AERPAW Testbed,”IEEE Communications Magazine, vol. 63, no. 2, pp. 78–84, 2025. doi: 10.1109/MCOM.001.2400327

work page doi:10.1109/mcom.001.2400327 2025

[2] [2]

NexRAN: Closed- loop RAN slicing in POWDER – A top-to-bottom open-source open-RAN use case,

D. Johnson, D. Maas, and J. Van Der Merwe, “NexRAN: Closed- loop RAN slicing in POWDER – A top-to-bottom open-source open-RAN use case,” inProc. WiNTECH, pp. 17–23, 2021. doi: 10.1145/3477086.3480842

work page arXiv 2021

[3] [3]

COSMOS: A city-scale programmable testbed for experimentation with advanced wireless,

D. Raychaudhuriet al., “COSMOS: A city-scale programmable testbed for experimentation with advanced wireless,” inProc. ACM MobiCom, pp. 1–13, 2020

work page 2020

[4] [4]

PandORA: Automated Design and Comprehensive Evaluation of Deep Reinforce- ment Learning Agents for Open RAN,

C. Tsampazi, M. Levorato, F. Restuccia, and T. Melodia, “PandORA: Automated Design and Comprehensive Evaluation of Deep Reinforce- ment Learning Agents for Open RAN,”IEEE Transactions on Mobile Computing, Early Access, 2025. Available: 10766614

work page 2025

[5] [5]

A Comparative Analysis of Deep Reinforcement Learning-Based xApps in O-RAN,

C. Tsampazi, F. Restuccia, and T. Melodia, “A Comparative Analysis of Deep Reinforcement Learning-Based xApps in O-RAN,” inProc. IEEE INFOCOM, 2023. Available: 10437367

work page 2023

[6] [6]

ColO- RAN: Developing Machine Learning-Based xApps for Open RAN Closed-Loop Control on Programmable Experimental Platforms,

M. Polese, L. Bonati, S. D’Oro, S. Sharma, and T. Melodia, “ColO- RAN: Developing Machine Learning-Based xApps for Open RAN Closed-Loop Control on Programmable Experimental Platforms,”IEEE Communications Magazine, 2022. Available: 9814869

work page 2022

[7] [7]

OrchestRAN: Orches- trating Network Intelligence in the Open RAN,

S. D’Oro, M. Polese, L. Bonati, and T. Melodia, “OrchestRAN: Orches- trating Network Intelligence in the Open RAN,”IEEE Communications Magazine, 2024. Available: 10356734

work page 2024

[8] [8]

REAL: Re- inforcement Learning-Enabled xApps for Experimental Closed-Loop Optimization in O-RAN with OSC RIC and srsRAN,

R. Barker, A. E. Dorcheh, T. Seyfi, and F. Afghah, “REAL: Re- inforcement Learning-Enabled xApps for Experimental Closed-Loop Optimization in O-RAN with OSC RIC and srsRAN,”arXiv preprint arXiv:2502.00715, 2025. Available: arXiv:2502.00715

arXiv 2025