Pith. sign in

REVIEW 5 major objections 4 minor 46 references

Access Probability Optimization in RACH: A Multi-Armed Bandits Approach

T0 review · 5 major / 4 minor · reviewed 2026-08-16 · deepseek-v4-flash

Pith's one-line read Multi-armed bandits can tune random-access preamble choices to favor high-priority users, reaching within 5% of the constrained optimum.

desk verdict A genuinely new non-uniform access scheme for two-priority RACH, with plausible results, but the headline claim that the MAB needs no knowledge of UE counts fails because the reward scaling in Section IV-C requires nh. read the letter →

arxiv 2504.14085 v1 pith:34PRWGBP submitted 2025-04-18 eess.SP

classification eess.SP
keywords massivemachine-typecommunications(mMTC)multi-armedbandits(MAB)randomaccesschannel(RACH)reinforcementlearning(RL)probabilityoptimizationnon-uniformpreambleselectionprioritynetworkloadestimation
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper's central claim is that non-uniform preamble selection in the cellular random access channel lets a base station favor high-priority devices while keeping low-priority service above a threshold, and that this choice can be learned online. It builds a two-priority model in which the base station sees each slot as a pattern of single high-priority successes, single low-priority successes, empty resource blocks, and collisions, and from these observable patterns it can score candidate access-probability vectors. The exact optimum is formulated as a constrained nonlinear problem, and a multi-armed bandit with a discretized action space is proposed to approximate it; the reported results put the bandit within 5% of the constrained optimum and show it adapting to a mid-run change in the number of devices. If the claims hold, the practical upshot is a scalable way to prioritize machine-type traffic without redesigning the RACH protocol.

What carries the argument

The load-bearing object is the access pattern: each RACH slot is summarized as a word over four event types—single high-priority success, single low-priority success, empty resource block, and collision—and expected throughputs are written as sums over the probabilities of these patterns. That model turns the problem into a constrained nonlinear optimization over the two probability vectors, and then into a multi-armed bandit in which each arm is a discretized pair of probability vectors, redundant circular shifts are removed offline, rewards are the observed high-priority throughput (penalized when low-priority throughput falls below the threshold), and cross-entropy updates concentrate the sampling distribution on high-reward arms without visiting every action.

What would settle it

Run the multi-armed bandit with the reward scaler replaced by a normalization computed only from observed throughputs (or set to 1), give the base station no access to the numbers of high- and low-priority devices, and compare the returned high-priority throughput with the sequential-quadratic-programming optimum for a fixed load such as $n_h=4$, $n_l=5$, $M=6$, $\gamma=0.4$; a gap beyond 5% or a violation of the low-priority constraint would refute the no-count claim as written.

Watch

Extended reading notes

Core claim

The paper's central discovery, stated in its own terms, is that the optimal preamble-selection probabilities in a two-priority RACH are non-uniform: high-priority UEs spread across most resource blocks while low-priority UEs concentrate on a few, and this allocation beats both uniform selection and access class barring on high-priority throughput under the same low-priority floor. The multi-armed bandit formulation reaches within 5% of the constrained optimum (Tables VII and VIII) and, unlike the sequential-quadratic-programming optimum, runs in roughly constant reported wall-clock time as the number of resource blocks grows from 3 to 6 in the constrained case; the compact action space variant cuts that time further and also produces a network-load estimate whose mean absolute error decreases with runs. The paper also reports that the bandit adapts when the network load changes mid-simulation.

Load-bearing premise

The reward-scaling step divides the measured high-priority throughput by a value computed from the numbers of high- and low-priority users, so the claim that the base station needs no knowledge of those counts holds only if a separate count or estimate is available.

Editorial extensions

If this is right

  • A base station can tune preamble selection probabilities for each priority class from observed access patterns alone, without changing the RACH protocol.
  • Non-uniform allocation gives higher high-priority throughput than uniform allocation or access class barring while holding low-priority throughput at the required threshold.
  • The multi-armed bandit evaluates only a small curated set of probability pairs, so the approach remains feasible when the number of resource blocks grows large enough to make exact optimization impractical.
  • The compact action space version can return an estimate of the number of high- and low-priority devices as a byproduct of the lookup table.
  • When the device population changes mid-operation, the learned probabilities re-converge to the new operating point.

Reading between the lines

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

  • The paper's conclusion lists 'the BS does not initially know the number of H-UEs and L-UEs' as future work, while the abstract claims no such knowledge is needed; this asymmetry suggests the count-free claim should be read as holding only after a load-estimation step has supplied the reward-scaling constant.
  • A straightforward testable extension is to replace the count-dependent reward scaler with a normalizer computed only from observed throughputs (for example, the batch maximum of $\mu_{h,T}$, or $M$ itself) and re-run Tables VII and VIII; if the within-5% gap survives, the no-count claim becomes empirically supported.
  • Because the model only needs the four event types, the same bandit framing should extend to more than two priority classes or to traffic with non-saturated arrivals, though the paper does not report those cases.
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 / 4 minor

Summary. The manuscript studies a two-priority random access channel (RACH) in which high-priority and low-priority UEs choose preambles according to non-uniform probability vectors. It defines a pattern-based throughput model, formulates a constrained optimization problem for the access probabilities, solves it with SQP, and then proposes a multi-armed bandit (MAB) approach with cross-entropy updates to approximate the solution. Two action-space constructions are considered: a discretized action space with circular-shift redundancy removal, and a compact action space (CAS) built from a precomputed lookup table. The paper's headline claim is that the MAB solution is near-optimal and scalable and does not require the BS to know the number of UEs.

Significance. The idea of optimizing non-uniform preamble-selection probabilities for prioritized RACH access is interesting, and the optimization results in Tables IV and V are plausible under the stated multinomial model. The MAB formulation with cross-entropy updates and the action-space reduction are useful engineering ideas. However, the central no-knowledge claim is not supported by the manuscript as written: the reward-scaling step requires knowledge of nh, the CAS evaluation is circular, and the simulation appendix is not reproducible as printed. If the no-knowledge claim could be substantiated, the contribution would be significant for mMTC deployment; in its current form, the paper needs substantial revision.

major comments (5)
  1. [Section IV-C, Algorithm 1, Eqs. (9)-(10)] The reward-scaling step divides µh,T by the value of µh obtained when ph and pl are set to (9)-(10). Under the multinomial model of Section II, this divisor equals nh(1 - 1/(M-1))^{nh-1}, which is an explicit function of the unknown nh. Algorithm 1 has no input for nh, and the experiments in Tables VII and VIII use known values nh=4, nl=5. A BS that does not know nh cannot evaluate the scaling factor, so the algorithm as specified is not deployable in the setting claimed in the abstract. The paper's own conclusion lists 'where the BS does not initially know the number of H-UEs and L-UEs' as future work, directly contradicting the abstract. Because scaling is applied inside every Q-value update, this is a load-bearing issue. Please either modify the scaling to be computable without nh, add an estimation stage, or remove the no-knowledge claim from the abstract and contributions.
  2. [Section IV-B2, Section V-D3] The compact action space is built by solving the same optimization problem from Section III-B for each (nh,nl) pair, and the numerical evaluation in Section V-D3 tests the MAB on the same load pairs that were used to populate the lookup table, including (nh,nl)=(4,5). The true optimum is therefore contained in the action set by construction, and the MAB is essentially selecting among precomputed optima rather than discovering near-optimal probabilities. This makes the reported CAS performance a tautology rather than evidence for generalization. Please test on held-out load pairs, or reposition the CAS as a table-lookup method with no generalization claim.
  3. [Section III-A, Eq. (3)] The probability P(π) is the foundation of the throughput metrics µh and µl in Eqs. (1)-(2) and of the optimization problem (4), but its derivation is deferred to the authors' own companion paper [34], and the recursive backtracking algorithm for enumerating Π is omitted. As a result, the optimization results cannot be verified from the manuscript alone. Please either include the derivation of the factors in Eq. (3) and the backtracking procedure, or clearly state that the model is inherited from [34] and provide enough detail for an independent check.
  4. [Appendix B, Algorithm 5] Algorithm 5 is not reproducible as printed. If C is the (nh+nl) by T matrix described in lines 3-5, then 'for c in C' iterates over rows (individual UEs), not over RACH slots, so the per-slot event labeling described in lines 12-21 is applied to per-UE time histories rather than to slot-wise access patterns. Furthermore, lines 14-21 identify whether a successful UE is H or L by comparing the slot index j with the RB index i, which is not meaningful; the H/L distinction should come from the row's position in C. Please provide the corrected pseudocode or the actual simulation code used to generate Tables II-VIII.
  5. [Table II] The uniform-access benchmark values in Table II appear inconsistent with the model of Section II. For nh=4, nl=5, and M=3, Eqs. (1)-(3) give µh = 4(2/3)^8 ≈ 0.156 and µl = 5(2/3)^8 ≈ 0.195, whereas Table II reports 0.31 and 0.62. For M=5 the model gives µh ≈ 0.671 and µl ≈ 0.839, versus the reported 0.40 and 0.53. Since Table II is used as the uniform benchmark in later comparisons, please reconcile these values with the model or the simulation, and clarify whether the reported numbers are means, medians, or values from a single realization.
minor comments (4)
  1. [Section IV-A, after Eq. (6)] The text says 'the reward is µl,T multiplied by the penalty factor' when the L-UE constraint is violated, but Eq. (6) defines the penalized reward as ρ µh,T. Please align the text with the equation.
  2. [Throughout] There are several typographical errors: 'Futhermore' in the Introduction, 'surpases' in the Conclusion, and 'utilitez' in Appendix A. The sentence about the simulation hardware is also duplicated in Section V.
  3. [Figures 3 and 4] In panels (d) of Figures 3 and 4, the legend shows 'µh,Tµh,T' instead of 'µh,Tµl,T'. Please correct the legends.
  4. [Section V-D4] The text reports a maximum achievable H-UE throughput of 1.2282 for nh=4, nl=5, but Table V gives µh = 1.28 for the same load at M=5. Please clarify whether 1.2282 is a scaled value or belongs to a different configuration.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation; the central MAB result is independent of the paper's inputs, though reward scaling has an implementability gap.

full rationale

The paper's claimed derivation chain is not circular. The throughput model (1)-(3) is a standard combinatorial probability calculation; the deferral of the detailed derivation of (3) to the authors' prior work [34] is a parameter-free mathematical identity and does not smuggle in the optimization result. The SQP optimization in Section III-B and the discrete-action-space MAB in Section IV use observed throughput counts (7)-(8) and are benchmarked against uniform allocation, ACB, and the SQP optimum (Tables II-VIII), so the near-optimality claim is externally testable rather than constructed. The CAS is explicitly presented as a lookup table of offline SQP solutions (Section IV-B2); reporting that the CAS MAB returns those precomputed optima is a retrieval result, not a derivation of the optimum from the MAB reward, and it is not used to justify the main MAB claim. The one significant defect is implementability, not circularity: Section IV-C defines the reward scaling factor as the value of mu_h under (9)-(10), which depends on nh, while the abstract and Algorithm 1 claim nh is unknown; Section VI even lists unknown nh/nl as future work. Because the scaling factor would be common to all actions, it does not change the argmax, so this is an internal inconsistency in the algorithm description rather than a circular reduction of the claimed prediction. No fitted parameter is renamed as a prediction, and no load-bearing argument reduces to a self-citation.

Assumptions & free parameters 10 free parameters · 6 assumptions · 0 invented entities

The central claim rests on a two-priority multinomial RACH model whose pattern probabilities are deferred to the authors' earlier work [34]. The MAB's evaluation uses many hand-picked hyperparameters. The reward scaling requires nh, the quantity the MAB claims not to need. The CAS lookup table is built from the same optimization problem and the same (nh,nl) values used in the tests, so the load-estimation byproduct is essentially a table readout rather than an independent estimator.

free parameters (10)
  • d = 0.2 (default)
    Discretization step for the action space; chosen empirically in Table VI; determines granularity and AS size.
  • rho = 0 (discrete AS), 0.1 (CAS)
    Penalty factor in reward (6); set per experiment; chosen empirically.
  • gamma = 0 or 0.4 in experiments
    L-UE throughput threshold in (4b); an operator-set constraint, not an output of the model.
  • alpha = 0.2 or 0.1
    Cross-entropy learning rate; chosen empirically in Tables VI and IX.
  • s = 0.1
    Elite fraction for CE update (Algorithm 2).
  • Bs = 500 or 200
    Batch size in Algorithm 1; chosen empirically.
  • T = 1000 or 100
    Number of RACH slots per reward evaluation in (7)-(8); chosen empirically.
  • runs = 15000, 30000, 2000, or 4000
    Total number of MAB iterations in experiments; chosen empirically.
  • N_h^s, N_l^s = 0 to 10 sweep
    Upper limits of the CAS lookup table; require a rough prior estimate of network load.
  • Reward scaling factor = depends on nh
    Section IV-C divides by µh from (9)-(10), a function of the unknown nh; not specified how the agent computes it.
assumptions (6)
  • domain assumption The probability of an access pattern factors as in (3), with factors derived in [34].
    Section III-A invokes this factorization to define µh and µl; the derivation is deferred to a self-cited unpublished reference.
  • domain assumption All UEs have a packet to transmit in every RACH slot.
    Section II states 'we assume all n = nh + nl UEs have a packet to transmit at each slot'; this full-buffer assumption drives the traffic model.
  • domain assumption Access patterns π1,...,πT are i.i.d. across slots.
    Section II: 'It is assumed that π1, π2,...,πT are independent and identically distributed (i.i.d.)'.
  • domain assumption The BS can perfectly detect single-H, single-L, empty, and collision events, but not the number of collided UEs.
    Section II event definitions; the observation model only distinguishes these four event types.
  • domain assumption SQP returns the global optimum of (4).
    Section III-B solves the non-convex constrained problem with SQP and labels the result 'optimal' without a global optimality certificate.
  • ad hoc to paper The reward scaling factor can be evaluated without knowledge of nh.
    Section IV-C defines scaling by µh under (9)-(10), which depends on nh; the paper does not supply a way to compute it from observations only, contradicting the no-knowledge claim.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Access Probability Optimization in RACH: A Multi-Armed Bandits Approach." pith.science (2026). https://pith.science/paper/34PRWGBP

@misc{pith2026250414085,
  author       = {Pith},
  title        = {Pith review of: Access Probability Optimization in RACH: A Multi-Armed Bandits Approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/34PRWGBP}},
  note         = {Machine review of arXiv:2504.14085}
}
read the original abstract

The use of cellular networks for massive machine-type communications (mMTC) is an appealing solution due to the availability of the existing infrastructure. However, the massive number of user equipments (UEs) poses a significant challenge to the cellular network's random access channel (RACH) regarding congestion and overloading. To mitigate this problem, we first present a novel approach to model a two-priority RACH, which allows us to define access patterns that describe the random access behavior of UEs as observed by the base station (BS). A non-uniform preamble selection scheme is proposed, offering increased flexibility in resource allocation for different UE priority classes. Then, we formulate an allocation model that finds the optimal access probabilities to maximize the success rate of high-priority UEs while constraining low-priority UEs. Finally, we develop a reinforcement learning approach to solving the optimization problem using multi-armed bandits, which provides a near-optimal but scalable solution and does not require the BS to know the number of UEs in the network.

Figures

Figures reproduced from arXiv: 2504.14085 by the authors.

Figure 1
Figure 1. Example of a pattern sequence over T time slots. chooses RB i, let p h = [p h 1 , . . . , ph M] and p l = [p l 1 , . . . , pl M] be vectors collecting the RB access probabilities for a H-UE, and an L-UE, respectively. Clearly, we have PM i=1 p h i = 1 and PM i=1 p l i = 1. We assume the access probability vectors p h and p l are known at the BS, which is responsible for generating and broadcasting these probabilitie… view at source ↗
Figure 2
Figure 2. shows µh,T and µl,T versus the number of runs for M = 6. We can see that the MAB formulation converges to the optimal solution for the unconstrained problem, with µh,T reaching 2.05 and µl,T reaching 0. Furthermore, the time taken to solve the optimization problem is not directly 0 5 10 15 0 0.5 1 1.5 2 Runs (×1000) µh,T , µl,T µh,T µl,T [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. µh,T and µl,T versus runs for γ = 0.4 and (3a) M = 3, (3b) M = 4, (3c) M = 5, (3d) M = 6. TABLE VIII MAB RESULTS WITH n h = 4, n l = 5, AND γ = 0.4. Output Variables Metrics M p h p l µh,T µl,T Time Taken (s) 3 [0, 0.2, 0.8] [0.2, 0, 0.8] 0.41 0.41 674.45 4 [0, 0.2, 0.6, 0.2] [0.2, 0, 0, 0.8] 0.82 0.41 670.74 5 [0, 0.2, 0.2, 0.2, 0.4] [0.2, 0, 0, 0, 0.8] 1.23 0.41 672.28 6 [0, 0.2, 0.2, 0.2, 0.2] [0.2, 0, 0, 0, 0, 0… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: µh,T and µl,T versus runs for γ = 0.4 and (3a) M = 3, (3b) M = 4, (3c) M = 5, (3d) M = 6 using the CAS. Algorithm 3 Discretization function with circular shift check 1: function DISCRETIZEASCS(M, r, s, E) 2: AS ← empty list 3: p h p = s[0], p l p = s[1] 4: if Pp h p = …
Figure 6
Figure 6. Figure 6: MAE versus runs for γ = 0.4 and M = 6. Algorithm 4 Check for Circular Shifts 1: function ISCIRCULARSHIFT(s, E) 2: CS ← empty list 3: for CS in range(|s[0]|) do 4: p h ← roll(|s[0]|, CS) 5: p l ← roll(|s[1]|, CS) 6: CS.append([P, X]) 7: for cs in CS do 8: for e in E do …
Figure 5
Figure 5. Figure 5: µh,T and µl,T versus runs for M = 5 under varying network load using (5a) the discretized AS and (5b) the CAS. The algorithm repeats this process for all combinations of p h and p l until the set AS is returned that contains all possible actions without circular shift …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

46 extracted references · 42 canonical work pages

  1. [34]

    Load estimation in a two-priority mMTC random access channel,

    A. O. Elmeligy, I. Psaromiligkos, and A. Minh, “Load estimation in a two-priority mMTC random access channel,” IEEE Veh. Technol. Conf. 2024 Fall, 2024, accepted for publication

  2. [1]

    Modeling and analysis of random access channels with bursty arrivals in OFDMA wireless networks,

    C.-H. Wei, G. Bianchi, and R.-G. Cheng, “Modeling and analysis of random access channels with bursty arrivals in OFDMA wireless networks,” IEEE Trans. on Wirel. Commun. , vol. 14, no. 4, pp. 1940– 1953, 2014

  3. [2]

    Is the random access channel of LTE and LTE-A suitable for m2m communications? a survey of alternatives,

    A. Laya, L. Alonso, and J. Alonso-Zarate, “Is the random access channel of LTE and LTE-A suitable for m2m communications? a survey of alternatives,” IEEE Commun. Surv. & Tutorials , vol. 16, no. 1, pp. 4–16, 2013

  4. [3]

    Toward 5G densenets: architectural advances for effective machine-type communications over femtocells,

    M. Condoluci, M. Dohler, G. Araniti, A. Molinaro, and K. Zheng, “Toward 5G densenets: architectural advances for effective machine-type communications over femtocells,” IEEE Commun. Mag. , vol. 53, no. 1, pp. 134–141, 2015

  5. [4]

    Hybrid random access and data transmission protocol for machine-to-machine communications in cellular networks,

    D. T. Wiriaatmadja and K. W. Choi, “Hybrid random access and data transmission protocol for machine-to-machine communications in cellular networks,” IEEE Trans. on Wirel. Commun. , vol. 14, no. 1, pp. 33–46, 2014

  6. [5]

    3GPP ts 37.869 v12.0.0 (2013-09),

    3rd Generation Partnership Project, “3GPP ts 37.869 v12.0.0 (2013-09),” Technical Specification Group Radio Access Network, Technical Specification 37.869, 2013, study on enhancements to Machine-Type Communications (MTC) and other mobile data applications; Radio Access Network (RAN) aspects. [Online]. Available: https://portal.3gpp.org/desktopmodules/ Spe...

  7. [6]

    How accurate is the RACH procedure model in LTE and LTE-A?

    O. Arouk, A. Ksentini, and T. Taleb, “How accurate is the RACH procedure model in LTE and LTE-A?” in 2016 International Wireless Communications and Mobile Computing Conference (IWCMC) . IEEE, 2016, pp. 61–66

  8. [7]

    A novel RACH scheme for efficient access in 5G and beyond networks using hash function,

    S. N. Swain and A. Subudhi, “A novel RACH scheme for efficient access in 5G and beyond networks using hash function,” in 2022 IEEE Future Networks World Forum (FNWF). IEEE, 2022, pp. 75–82

Show all 46 references
  1. [8]

    Random access channel management for handling massive numbers of machine-to-machine communication devices,

    A. D. A. Kurnia and Y .-J. Choi, “Random access channel management for handling massive numbers of machine-to-machine communication devices,” in 2018 International Conference on Information and Commu- nication Technology Convergence (ICTC). IEEE, 2018, pp. 1186–1190

  2. [9]

    Statistical priority-based multiple access system and method,

    S. M. Clark, K. A. Hoback, and S. J. F. Zogg, “Statistical priority-based multiple access system and method,” Mar. 2010. [Online]. Available: https://patents.google.com/patent/US7680077B1/en

  3. [10]

    Modeling and performance analysis of statistical priority-based multiple access protocol,

    M. Yang, Q. Zhang, Y . Tu, and H. Yu, “Modeling and performance analysis of statistical priority-based multiple access protocol,” in 2018 IEEE/CIC International Conference on Communications in China (ICCC). IEEE, 2018, pp. 399–404

  4. [11]

    Performance analysis of the statistical priority-based multiple access,

    J. Liu, T. Peng, Q. Quan, and L. Cao, “Performance analysis of the statistical priority-based multiple access,” in 2017 3rd IEEE International Conference on Computer and Communications (ICCC) , Dec. 2017, pp. 30–35

  5. [12]

    Modeling and performance analysis of statistical priority-based multiple access: A stochastic geometry approach,

    Y . Zhang, Y . He, X. Wang, H. Sun, and T. Q. S. Quek, “Modeling and performance analysis of statistical priority-based multiple access: A stochastic geometry approach,” IEEE Internet Things J., vol. 9, no. 15, pp. 13 942–13 954, Aug. 2022, conference Name: IEEE Internet of Th...

  6. [13]

    Priority- based load-adaptive preamble separation random access for QoS- differentiated services in 5G networks,

    H. Althumali, M. Othman, N. K. Noordin, and Z. M. Hanapi, “Priority- based load-adaptive preamble separation random access for QoS- differentiated services in 5G networks,” J. Netw. Comput. Appl. , vol. 203, p. 103396, 2022

  7. [14]

    D-ACB: Adaptive congestion control algorithm for bursty M2M traffic in LTE networks,

    S. Duan, V . Shah-Mansouri, Z. Wang, and V . W. Wong, “D-ACB: Adaptive congestion control algorithm for bursty M2M traffic in LTE networks,” IEEE Trans. on Veh. Technol., vol. 65, no. 12, pp. 9847–9861, 2016

  8. [15]

    Queue-aware access prioritization for massive machine-type communication,

    M. R. Chowdhury and S. De, “Queue-aware access prioritization for massive machine-type communication,” IEEE Internet Things J. , vol. 9, no. 17, pp. 15 858–15 873, Sep. 2022, conference Name: IEEE Internet of Things Journal

  9. [16]

    Online control of preamble groups with priority in massive iot networks,

    J. Liu, M. Agiwal, M. Qu, and H. Jin, “Online control of preamble groups with priority in massive iot networks,” IEEE J. on Sel. Areas Commun., vol. 39, no. 3, pp. 700–713, 2020

  10. [17]

    Utilizing cognitive radio and reinforcement learning for multiple preamble detection in 5G- PRACH,

    S. S. Zehra, R. Qureshi, M. Aamir, and M. Magarini, “Utilizing cognitive radio and reinforcement learning for multiple preamble detection in 5G- PRACH,” in 2022 25th International Symposium on Wireless Personal Multimedia Communications (WPMC) . IEEE, 2022, pp. 70–74

  11. [18]

    Deep reinforcement learning mechanism for dynamic access control in wireless networks handling mMTC,

    D. Pacheco-Paramo, L. Tello-Oquendo, V . Pla, and J. Martinez-Bauset, “Deep reinforcement learning mechanism for dynamic access control in wireless networks handling mMTC,” Ad Hoc Networks , vol. 94, p. 101939, 2019

  12. [19]

    A reinforcement-learning- based access scheme for low-latency and correlated-traffic mtc networks,

    D. T. Nguyen, X. Zhan, and T. Le-Ngoc, “A reinforcement-learning- based access scheme for low-latency and correlated-traffic mtc networks,” in 2022 IEEE Wireless Communications and Networking Conference (WCNC). IEEE, 2022, pp. 620–625

  13. [20]

    Reinforcement learning based dynamic resource allocation for massive MTC in sliced mobile networks,

    B. Yang, Y . Xu, X. She, J. Zhu, F. Wei, P. Cheri, and J. Wang, “Reinforcement learning based dynamic resource allocation for massive MTC in sliced mobile networks,” in 2022 IEEE 14th International Conference on Advanced Infocomm Technology (ICAIT) . IEEE, 2022, pp. 298–303

  14. [21]

    Reinforcement learning based preamble resource allocation scheme for access control in machine-to-machine communication,

    H. Liu, B. Liu, H. Gao, X. Xu, and X. Su, “Reinforcement learning based preamble resource allocation scheme for access control in machine-to-machine communication,” in International Conference on Communications and Networking in China . Springer, 2022, pp. 44–56

  15. [22]

    Priority- based learning automata in Q-learning random access scheme for cellular M2M communications,

    N. A. Shinkafi, L. M. Bello, D. S. Shu’aibu, and P. D. Mitchell, “Priority- based learning automata in Q-learning random access scheme for cellular M2M communications,” ETRI J., vol. 43, no. 5, pp. 787–798, 2021

  16. [23]

    PPO-based PDACB traffic control scheme for massive IoV communications,

    H. Zhang, M. Jiang, X. Liu, X. Wen, N. Wang, and K. Long, “PPO-based PDACB traffic control scheme for massive IoV communications,” IEEE Trans. on Intell. Transp. Syst. , 2022

  17. [24]

    Deep reinforcement learning based flexible preamble allocation for RAN slicing in 5G networks,

    A. M. Gedikli, M. Koseoglu, and S. Sen, “Deep reinforcement learning based flexible preamble allocation for RAN slicing in 5G networks,” Comput. Networks, vol. 215, p. 109202, 2022

  18. [25]

    Prioritized random access with dynamic access barring for RAN overload in 3GPP LTE-A networks,

    J.-P. Cheng, C.-h. Lee, and T.-M. Lin, “Prioritized random access with dynamic access barring for RAN overload in 3GPP LTE-A networks,” in 2011 IEEE GLOBECOM Workshops (GC Wkshps) . IEEE, 2011, pp. 368–372. 14

  19. [26]

    PRADA: Prioritized random access with dynamic access barring for MTC in 3GPP LTE-A networks,

    T.-M. Lin, C.-H. Lee, J.-P. Cheng, and W.-T. Chen, “PRADA: Prioritized random access with dynamic access barring for MTC in 3GPP LTE-A networks,” IEEE Trans. on Veh. Technol. , vol. 63, no. 5, pp. 2467– 2472, Jun. 2014, conference Name: IEEE Transactions on Vehicular Technology

  20. [27]

    Handling mission- critical communication in smart grid distribution automation services through lte,

    C. Kalalas, F. Vazquez-Gallego, and J. Alonso-Zarate, “Handling mission- critical communication in smart grid distribution automation services through lte,” in 2016 IEEE International Conference on Smart Grid Communications (SmartGridComm). IEEE, 2016, pp. 399–404

  21. [28]

    Multiagent reinforcement learning meets random access in massive cellular internet of things,

    J. Bai, H. Song, Y . Yi, and L. Liu, “Multiagent reinforcement learning meets random access in massive cellular internet of things,” IEEE Internet Things J., vol. 8, no. 24, pp. 17 417–17 428, 2021

  22. [29]

    Random access scheme for machine type communication networks using reinforcement learning approach,

    P. Orim, N. Ventura, and J. Mwangama, “Random access scheme for machine type communication networks using reinforcement learning approach,” in 2023 IEEE AFRICON . IEEE, 2023, pp. 1–6

  23. [30]

    A minimalist approach to offline reinforcement learning,

    S. Fujimoto and S. S. Gu, “A minimalist approach to offline reinforcement learning,” Adv. Neural Inf. Process. Syst. , vol. 34, pp. 20 132–20 145, 2021

  24. [31]

    Fast adaptive S-ALOHA scheme for event-driven machine-to-machine communications,

    H. Wu, C. Zhu, R. J. La, X. Liu, and Y . Zhang, “Fast adaptive S-ALOHA scheme for event-driven machine-to-machine communications,” in 2012 IEEE vehicular technology conference (VTC Fall) . IEEE, 2012, pp. 1–5

  25. [32]

    Stabilizing multi-channel slotted aloha for machine-type communications,

    O. Galinina, A. Turlikov, S. Andreev, and Y . Koucheryavy, “Stabilizing multi-channel slotted aloha for machine-type communications,” in 2013 IEEE International Symposium on Information Theory . IEEE, 2013, pp. 2119–2123

  26. [33]

    Optimal access class barring for stationary machine type communication devices with timing advance information,

    Z. Wang and V . W. Wong, “Optimal access class barring for stationary machine type communication devices with timing advance information,” IEEE Trans. on Wirel. Commun. , vol. 14, no. 10, pp. 5374–5387, 2015

  27. [35]

    NLPQL: A FORTRAN subroutine solving constrained nonlinear programming problems,

    K. Schittkowski, “NLPQL: A FORTRAN subroutine solving constrained nonlinear programming problems,” Ann. Oper. Res. , vol. 5, pp. 485–500, 1986

  28. [36]

    Finite-time analysis of the multiarmed bandit problem,

    P. Auer, N. Cesa-Bianchi, and P. Fischer, “Finite-time analysis of the multiarmed bandit problem,” Mach. learning, vol. 47, pp. 235–256, 2002

  29. [37]

    A survey of online exper- iment design with the stochastic multi-armed bandit,

    G. Burtini, J. Loeppky, and R. Lawrence, “A survey of online exper- iment design with the stochastic multi-armed bandit,” arXiv preprint arXiv:1510.00757, 2015

  30. [38]

    Internally re- warded reinforcement learning,

    M. Li, X. Zhao, J. H. Lee, C. Weber, and S. Wermter, “Internally re- warded reinforcement learning,” in International Conference on Machine Learning. PMLR, 2023, pp. 20 556–20 574

  31. [39]

    Learning more skills through optimistic exploration,

    D. Strouse, K. Baumli, D. Warde-Farley, V . Mnih, and S. Hansen, “Learning more skills through optimistic exploration,” arXiv preprint arXiv:2107.14226, 2021

  32. [40]

    Reward scale robustness for proximal policy optimization via dreamerv3 tricks,

    R. Sullivan, A. Kumar, S. Huang, J. Dickerson, and J. Suarez, “Reward scale robustness for proximal policy optimization via dreamerv3 tricks,” Adv. Neural Inf. Process. Syst. , vol. 36, 2024

  33. [41]

    Playing atari with deep reinforcement learning,

    V . Mnih, K. Kavukcuoglu, D. Silver, A. Graves, I. Antonoglou, D. Wier- stra, and M. Riedmiller, “Playing atari with deep reinforcement learning,” arXiv preprint arXiv:1312.5602 , 2013

  34. [42]

    Human-level control through deep reinforcement learning,

    V . Mnih, K. Kavukcuoglu, D. Silver, A. A. Rusu, J. Veness, M. G. Bellemare, A. Graves, M. Riedmiller, A. K. Fidjeland, G. Ostrovski et al., “Human-level control through deep reinforcement learning,” nature, vol. 518, no. 7540, pp. 529–533, 2015

  35. [43]

    A generative adversarial network enabled deep distributional reinforcement learning for transmis- sion scheduling in internet of vehicles,

    F. Naeem, S. Seifollahi, Z. Zhou, and M. Tariq, “A generative adversarial network enabled deep distributional reinforcement learning for transmis- sion scheduling in internet of vehicles,” IEEE Trans. on Intell. Transp. Syst., vol. 22, no. 7, pp. 4550–4559, 2020

  36. [44]

    Cemab: A cross-entropy- based method for large-scale multi-armed bandits,

    E. Wang, H. Kurniawati, and D. P. Kroese, “Cemab: A cross-entropy- based method for large-scale multi-armed bandits,” in Artificial Life and Computational Intelligence: Third Australasian Conference, ACALCI 2017, Geelong, VIC, Australia, January 31–February 2, 2017, Proceed- i...

  37. [45]

    R. Y . Rubinstein and D. P. Kroese, The cross-entropy method: a unified approach to combinatorial optimization, Monte-Carlo simulation, and machine learning. Springer, 2004, vol. 133

  38. [46]

    Study on ran improvements for machine type communications,

    3GPP, “Study on ran improvements for machine type communications,” TR 37.868, Tech. Rep., 2011. Ahmed O. Elmeligy Ahmed O. Elmeligy received the B.Sc. and M.Sc. degrees in Electrical Engineering from the American University of Sharjah, Sharjah, United Arab Emirates, in 2020 an...

Pith tools

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