Pith. sign in

REVIEW 1 major objections 6 minor 1 cited by

Enhancing Disaster Resilience with UAV-Assisted Edge Computing: A Reinforcement Learning Approach to Managing Heterogeneous Edge Devices

T0 review · 1 major / 6 minor · reviewed 2026-08-10 · deepseek-v4-flash

Pith's one-line read A drone-mounted scheduler learns which disaster sensors will fail first and extends their working life.

desk verdict A plausible, re-implementable simulation study of DQN for UAV edge computing under disasters, but the headline claims about learning to prioritize traffic-critical devices are not yet supported by the experiments as presented. read the letter →

arxiv 2501.15305 v1 pith:NBNTDKXA submitted 2025-01-25 cs.ET cs.AIcs.DC

classification cs.ETcs.AIcs.DC
keywords UAV-assistededgecomputingdisasterresiliencedeepQ-networkreinforcementlearningtaskoffloadingdevicelifetimedataageevacuationsimulation
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 claims that a single UAV, flown by a deep Q-network policy, can stretch the operating lifetime of heterogeneous edge-computing devices during disasters where some devices lose power and others lose communication. The UAV visits devices to offload their computations and relay their data, and the learned policy balances battery drain against the age of uncommunicated data. The authors report that the policy identifies the device most likely to fail first in a given configuration, giving maintenance crews a concrete priority target. They also show, in simulated evacuations of a rural town and an urban downtown, that devices near high-traffic roads are kept from failing first even when power and connectivity outages are randomized.

What carries the argument

The machinery is a deep Q-network scheduler whose state is a vector of remaining battery fractions and data ages for all devices, whose action is choosing which device the UAV visits next, and whose reward adds UAV time steps, the log ratio of chosen-device data age to oldest data age, and battery and age terms. The environment is a Gym simulation where each time slot consists of the UAV flying to a device, offloading computation, and all other devices processing locally, with Shannon's formula giving the transmission rate. In the evacuation cases, the reward is augmented in proportion to the traffic density near each device, which is what steers priority toward high-flow roads.

What would settle it

Run the same deep Q-network training in a physical testbed or in a second, independently written simulator with the same device types and outage configurations, and compare the first-failing device and episode lengths; if the predicted critical device differs in a substantial fraction of configurations, the simulator-grounded claim fails to transfer.

Watch

Extended reading notes

Core claim

In the paper's own terms, the discovery is that deep Q-network reinforcement learning can learn an effective UAV scheduling policy for a disaster-stricken edge network without a model of the disaster dynamics, under combined power and communication failures. The system maximizes the number of time slots before any device fails, where failure is defined as battery exhaustion or data age exceeding a limit, and in doing so it identifies the most critical device. In the evacuation simulations, the learned policy reliably protects devices on high-density evacuation routes: none of the rural high-traffic devices failed first in 30 randomized tests, and the urban high-traffic devices were likewise shielded despite occasional anomalies.

Load-bearing premise

The policy's success depends on the simulator's dynamics, including the imported power and channel parameters and the reward function chosen from the same simulator, matching real disaster conditions closely enough that a trained policy keeps working outside the simulation.

Editorial extensions

If this is right

  • If the policy works as modeled, operators could use the first-failing device prediction to dispatch maintenance to the single most critical device during an outage.
  • The episode-length results give a quantitative map of how network lifetime degrades as the number of devices with power and communication availability drops from 12 to 4.
  • The evacuation results imply that traffic-density-weighted rewards can embed domain knowledge into the scheduler without changing the reinforcement learning algorithm.
  • Since the state is just battery fractions and data ages, the same deep Q-network approach can be resized to any number of devices by enlarging the state vector.
  • Multiple-UAV deployment, flagged as future work, is a natural next step because single-UAV lifetime is ultimately capped by the UAV battery.

Reading between the lines

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

  • Beyond the paper's claims, the reward using the ratio of selected-device data age to oldest data age suggests a broader principle: prioritizing the device with the highest urgency gap, a rule that could transfer to other scheduling domains.
  • An implicit testable extension is to calibrate the imported channel and hardware parameters with real UAV-device measurements and re-run the same simulator; large deviations in per-device lifetime would indicate where the model departs from reality.
  • The anomalies in the evacuation results hint that the policy's protection is statistical rather than guaranteed, and confidence intervals over many random configurations would characterize that uncertainty more precisely.
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

1 major / 6 minor

Summary. The paper studies a UAV-assisted mobile edge computing system for disaster scenarios in which heterogeneous edge devices (e.g., Raspberry Pi, Jetson Nano) may lose power supply or communication connectivity. The authors formulate an optimization problem over UAV routing that trades off device battery consumption against the age of uncommunicated data, and then train a deep Q-network (DQN) to select which device to visit in each time slot. The system is evaluated in synthetic scenarios with varying numbers of powered and connected devices, and in two SUMO-based evacuation case studies (rural Round Lake and urban Albany). The headline claims are that the learned policy extends the operational lifetime of the network and identifies the device most likely to fail first, and that in the evacuation scenarios it learns to prioritize devices near high-traffic roads.

Significance. If the results are correct, the paper would make a modest contribution: it adds power and communication outage constraints and data-age constraints to a well-studied UAV-assisted MEC problem, and demonstrates a DQN-based scheduler in realistic traffic-evacuation settings. The use of established hardware parameters from prior work (Tables 2 and 3) and the connection to SUMO traffic simulation are strengths. However, the formal problem formulation contains material errors, and the central behavioral claim about traffic-aware prioritization is not supported by the experimental design because no baseline policy is tested. The reward function is also selected on the same episode-length metric that constitutes the main evaluation, which weakens the persuasiveness of the reported lifetime improvements.

major comments (1)
  1. [Section 5.1, Table 7] The claim that "the UAV successfully learns to prioritize devices close to high traffic flow" is not supported by the experiments as presented. Table 7 reports failure counts for the trained DQN only; there is no random-action, round-robin, nearest-device, or greedy baseline under the same 30 random power/communication configurations. The observed pattern could arise from the explicit traffic-density bonus in the reward function, from incidental correlations between device location and which devices are assigned outages, or from asymmetries in battery capacity and processing rates. The authors should add control policies and report the same failure-count table for them; this is necessary to attribute the prioritization to the learned policy.
minor comments (6)
  1. [Title] The title contains a typo: "UA V-Assisted" should be "UAV-Assisted".
  2. [Section 2] The sentence "since the the different optimization problems are non-convex" has a duplicated article; it should read "since the different optimization problems are non-convex".
  3. [Section 3.2] The description of the reward function is ambiguous. The text first gives a generic example r_t = log(A_t)+log(U_t)+log(M_t)-log(O_t), then states that the chosen reward is "the sum of the number of time slots elapsed and the log of the ratio of the age of data in the chosen device to the oldest data in any device." The reader must infer that this is U + log(A/O); please state the exact functional form in a display equation and define all symbols consistently.
  4. [Section 4, Eq. (4)] The noise power is listed as -100 dBm in Table 1, but in Eq. (4) it is denoted sigma^2. dBm is a unit of power, not power squared; please clarify the units and whether sigma^2 is the noise power or the noise variance.
  5. [Section 5] Table 5 reports mean episode lengths without any measure of variance, and Figure 3b shows that initial conditions strongly affect the maximum possible episode length. Reporting standard deviations or confidence intervals across seeds would strengthen the quantitative claims.
  6. [Section 5.1] The explanation of the traffic-density reward is brief: it says the RL system was "programmed to receive additional rewards if an edge device was situated near a high-density road, in proportion to the observed average density." Please specify how the density is computed, how the proportionality is calibrated, and whether this shaping is included in the final reward used for the results in Table 7.

Circularity Check

2 steps flagged · score 6.0 of 10

The evacuation-prioritization conclusion restates a reward-shaping term, and the reported lifetime result is the same episode-length metric used to select the final reward.

  1. self definitional [Section 5.1 (Prioritizing the Evacuation Route), page 14]
    "The UAV-edge RL system was programmed to receive additional rewards if an edge device was situated near a high-density road, in proportion to the observed average density. ... These results indicate that the UAV successfully learns to prioritize devices close to high traffic flow despite variations in power and communication constraints."

    The claimed outcome is written directly into the reward function: the agent is given extra reward for serving devices near high-density roads. Maximizing cumulative reward therefore gives the policy a direct incentive to protect those devices, so the observation that they fail less often is a restatement of the reward design rather than an independently discovered prioritization. The paper's own acknowledgment of anomalies and its closing comment that 'there is room for further exploration especially in the design of the reward function' further shows that the result is tied to the chosen reward rather than an external ground truth.

  2. fitted input called prediction [Section 4 (Implementation) and Section 5, Tables 5 and 6]
    "The most consistently favorable reward was found to be the sum of the number of time slots elapsed and the log of the ratio of the age of data in the chosen device to the oldest data in any device. ... the deep neural network will learn this over hundreds of thousands of iterations and take the action that is most likely to maximize the episode length. ... while the final reward listed in the table performs the best in combination with DQN across all scenarios."

    The reward function is chosen by comparing Table 6 entries, which are maximum episode lengths, and the paper then reports average episode length in Table 5 as the demonstrated lifetime extension of the system. Since the selection metric and the reported result are the same quantity, the headline lifetime figures partly reflect the reward-selection process rather than an independent evaluation on a held-out objective. No external utility or separate validation set is introduced to break the loop.

full rationale

Most of the simulation pipeline is not circular: the DQN is trained in a Gym environment with device, task, and channel parameters taken from independent prior work, the optimization objective is stated before training, and the basic observation that episode length improves with training is a genuine simulation finding. The two circular elements are localized and partial. The evacuation claim reduces to reward shaping because 'high-traffic prioritization' is inserted into the reward rather than discovered; the reward-function comparison selects the final reward using the same episode-length metric later presented as the main result. There are no load-bearing self-citations: the authors' own prior work appears only as an application example, and the imported physics and hardware parameters are externally sourced. The absence of a random or greedy control policy in Table 7 is a real experimental weakness, but it is a correctness concern rather than circularity. Overall, the core RL framework remains independently specified, but the two identified steps make part of the claimed demonstration self-referential, giving a partial circularity score of 6.

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

The reported lifetime gains rest on a hand-picked reward function, simulation parameters borrowed from prior work, and uniform random scenario generation. The paper introduces no fundamentally new entity; the UAV and edge devices are existing technologies.

free parameters (4)
  • Final reward function = U + log(A/O) from Table 6
    Selected post hoc as the best of 10 candidates on the same episode-length metric used for the headline results (Table 6, Section 5).
  • DQN learning rate = 0.0071
    Tuned down from 0.01 to obtain stable convergence; the paper states hyperparameter tuning was required (Section 5).
  • Traffic-density reward weight = unspecified
    Section 5.1 says additional rewards are added in proportion to observed average road density, but the proportionality constant is not reported, so the evacuation results depend on an unquantified choice.
  • Simulation region size = 800 m by 800 m
    Chosen by trial and error so that UAV range is not the limiting factor (Section 5); different regions would change episode lengths.
assumptions (4)
  • domain assumption DQN with the stated hyperparameters converges to a good policy in the simulated MDP.
    Convergence is asserted from training curves (Figure 3) without proof or a held-out evaluation.
  • domain assumption Power consumption and processing rates from [24] and the channel model from [3] accurately represent the simulated devices.
    Tables 1 through 3 import hardware and channel parameters from prior literature without calibration to the devices or disaster setting (Section 4).
  • ad hoc to paper All random parameters (device locations, types, tasks, battery capacities, and outage assignments) are uniformly distributed.
    The paper assumes uniform distributions for every random parameter (Section 3.1); real failure patterns may not be uniform.
  • domain assumption A visited device's data age resets to zero immediately, regardless of the volume of offloaded data or UAV computation time.
    Equation (1f) sets D_j(t+1) to zero whenever the UAV visits device j; UAV computation time is treated as negligible.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Enhancing Disaster Resilience with UAV-Assisted Edge Computing: A Reinforcement Learning Approach to Managing Heterogeneous Edge Devices." pith.science (2026). https://pith.science/paper/NBNTDKXA

@misc{pith2026250115305,
  author       = {Pith},
  title        = {Pith review of: Enhancing Disaster Resilience with UAV-Assisted Edge Computing: A Reinforcement Learning Approach to Managing Heterogeneous Edge Devices},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NBNTDKXA}},
  note         = {Machine review of arXiv:2501.15305}
}
read the original abstract

Edge sensing and computing is rapidly becoming part of intelligent infrastructure architecture leading to operational reliance on such systems in disaster or emergency situations. In such scenarios there is a high chance of power supply failure due to power grid issues, and communication system issues due to base stations losing power or being damaged by the elements, e.g., flooding, wildfires etc. Mobile edge computing in the form of unmanned aerial vehicles (UAVs) has been proposed to provide computation offloading from these devices to conserve their battery, while the use of UAVs as relay network nodes has also been investigated previously. This paper considers the use of UAVs with further constraints on power and connectivity to prolong the life of the network while also ensuring that the data is received from the edge nodes in a timely manner. Reinforcement learning is used to investigate numerous scenarios of various levels of power and communication failure. This approach is able to identify the device most likely to fail in a given scenario, thus providing priority guidance for maintenance personnel. The evacuations of a rural town and urban downtown area are also simulated to demonstrate the effectiveness of the approach at extending the life of the most critical edge devices.

Figures

Figures reproduced from arXiv: 2501.15305 by the authors.

Figure 1
Figure 1. Proposed System overview of time slots of the UAV subject to some constraints. Formally, 𝑥 𝑡 𝑖𝑗 is a binary variable that is 1 if the UAV moves from node 𝑖 to node 𝑗 at time slot 𝑡, and 0 otherwise. This information is formulated in statements (1a) and (1b): max Í 𝑡 Í 𝑖 Í 𝑗 𝑥 𝑡 𝑖𝑗 (1a) s.t. 𝑥 𝑡 𝑖𝑗 ∈ {0, 1} ∀𝑖, 𝑗, 𝑡 (1b) Í 𝑗 𝑥 𝑡+1 𝑖𝑗 = Í 𝑗 𝑥 𝑡 𝑗𝑖 ∀𝑖, 𝑡 (1c) 𝐸𝑗 (𝑡 + 1) = 𝐸𝑗 (𝑡) − 𝛿𝐸𝑗 hÍ 𝑖 𝑥 𝑡 𝑖𝑗 𝐸offload (𝑗) − (1 − Í … view at source ↗
Figure 2
Figure 2. Example layout of edge devices in the region with the start position of the UAV at the origin. Distance is in meters, [PITH_FULL_IMAGE:figures/full_fig_p010_2.png] view at source ↗
Figure 3
Figure 3. Reinforcement learning training showing the increase in average episode length as the DQN learns the best policy. [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Road network with numbered edge devices. Traffic density on the roads is color coded in order of increasing density: [PITH_FULL_IMAGE:figures/full_fig_p013_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

  1. Task Assignment and Exploration Optimization for Low Altitude UAV Rescue via Generative AI Enhanced Multi-agent Reinforcement Learning

    cs.AI 2025-04 reject novelty 4.0 of 10

    A Hungarian-plus-diffusion multi-agent RL controller with a Lyapunov energy queue assigns UAV rescue tasks to ground robots and airships, reporting lower latency and steadier energy use in simulation.

Reference graph

Works this paper leans on

34 extracted references · 32 canonical work pages · cited by 1 Pith paper

  1. [1]

    Mohamed Aboualola, Khalid Abualsaud, Tamer Khattab, Nizar Zorba, and Hossam S Hassanein. 2023. Edge technologies for disaster management: A survey of social media and artificial intelligence integration. IEEE Access (2023)

  2. [2]

    Talha Azfar, Chengyue Wang, Ruimin Ke, Adeeba Raheem, Jeffrey Weidner, and Ruey L Cheu. 2023. Incorporating Vehicle Detection Algorithms via Edge Computing on a Campus Digital Twin Model. In International Conference on Transportation and Development 2023 . 400–409

  3. [3]

    Xu Chen, Lei Jiao, Wenzhong Li, and Xiaoming Fu. 2015. Efficient multi-user computation offloading for mobile-edge cloud computing. IEEE/ACM transactions on networking 24, 5 (2015), 2795–2808

  4. [4]

    Milan Erdelj, Enrico Natalizio, Kaushik R Chowdhury, and Ian F Akyildiz. 2017. Help from the sky: Leveraging UAVs for disaster management. IEEE Pervasive Computing 16, 1 (2017), 24–32

  5. [5]

    Teruo Higashino, Hirozumi Yamaguchi, Akihito Hiromori, Akira Uchiyama, and Keiichi Yasumoto. 2017. Edge computing and IoT based research for building safe smart cities resistant to disasters. In 2017 IEEE 37th international conference on distributed computing systems (ICDCS). IEEE, 1729–1737

  6. [6]

    Razin Farhan Hussain, Mohsen Amini Salehi, Anna Kovalenko, Yin Feng, and Omid Semiari. 2019. Federated edge computing for disaster management in remote smart oil fields. In 2019 IEEE 21st International Conference on High Performance Computing and Communications; IEEE 17th International Conference on Smart City . IEEE, 929–936

  7. [7]

    Zeeshan Kaleem, Muhammad Yousaf, Aamir Qamar, Ayaz Ahmad, Trung Q Duong, Wan Choi, and Abbas Jamalipour. 2019. UAV- empowered disaster-resilient edge architecture for delay-sensitive communication. IEEE Network 33, 6 (2019), 124–132

  8. [8]

    Daniel Krajzewicz, Georg Hertkorn, Christian Rössel, and Peter Wagner. 2002. SUMO (Simulation of Urban MObility)-an open-source traffic simulation. In Proceedings of the 4th middle East Symposium on Simulation and Modelling (MESM20002) . 183–187. ACM J. Auton. Transport. Syst., Vol. 37, No. 4, Article 111. Publication date: November 2024. 111:16 • Talha A...

Show all 34 references
  1. [9]

    Mushu Li, Nan Cheng, Jie Gao, Yinlu Wang, Lian Zhao, and Xuemin Shen. 2020. Energy-efficient UAV-assisted mobile edge computing: Resource allocation and trajectory optimization. IEEE Transactions on Vehicular Technology 69, 3 (2020), 3424–3438

  2. [10]

    Fang Liu, Yeting Guo, Zhiping Cai, Nong Xiao, and Ziming Zhao. 2019. Edge-enabled disaster rescue: a case study of searching for missing people. ACM Transactions on Intelligent Systems and Technology (TIST) 10, 6 (2019), 1–21

  3. [11]

    Guanxiong Liu, Hang Shi, Abbas Kiani, Abdallah Khreishah, Joyoung Lee, Nirwan Ansari, Chengjun Liu, and Mustafa Mohammad Yousef. 2021. Smart traffic monitoring system using computer vision and edge computing. IEEE Transactions on Intelligent Transportation Systems 23, 8 (2021)...

  4. [12]

    Qian Liu, Long Shi, Linlin Sun, Jun Li, Ming Ding, and Feng Shu. 2020. Path planning for UAV-mounted mobile edge computing with deep reinforcement learning. IEEE Transactions on Vehicular Technology 69, 5 (2020), 5723–5728

  5. [13]

    Yizhe Luo, Wenrui Ding, and Baochang Zhang. 2021. Optimization of task scheduling and dynamic service strategy for multi-UAV-enabled mobile-edge computing system. IEEE Transactions on Cognitive Communications and Networking 7, 3 (2021), 970–984

  6. [14]

    Abbas Mehrabi and Kiseon Kim. 2015. Maximizing data collection throughput on a path in energy harvesting sensor networks using a mobile sink. IEEE Transactions on Mobile Computing 15, 3 (2015), 690–704

  7. [15]

    Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Andrei A Rusu, Joel Veness, Marc G Bellemare, Alex Graves, Martin Riedmiller, Andreas K Fidjeland, Georg Ostrovski, et al. 2015. Human-level control through deep reinforcement learning. nature 518, 7540 (2015), 529–533

  8. [16]

    Syed Ahsan Raza Naqvi, Syed Ali Hassan, Haris Pervaiz, and Qiang Ni. 2018. Drone-aided communication as a key enabler for 5G and resilient public safety networks. IEEE Communications Magazine 56, 1 (2018), 36–42

  9. [17]

    Antonin Raffin, Ashley Hill, Adam Gleave, Anssi Kanervisto, Maximilian Ernestus, and Noah Dormann. 2021. Stable-baselines3: Reliable reinforcement learning implementations. Journal of Machine Learning Research 22, 268 (2021), 1–8

  10. [18]

    Moataz Samir, Sanaa Sharafeddine, Chadi M Assi, Tri Minh Nguyen, and Ali Ghrayeb. 2019. UAV trajectory planning for data collection from time-constrained IoT devices. IEEE Transactions on Wireless Communications 19, 1 (2019), 34–46

  11. [19]

    Zaiba Shah, Umer Javed, Muhammad Naeem, Sherali Zeadally, and Waleed Ejaz. 2023. Mobile edge computing (MEC)-enabled UAV placement and computation efficiency maximization in disaster scenario. IEEE Transactions on Vehicular Technology 72, 10 (2023), 13406–13416

  12. [20]

    Claude Elwood Shannon. 1949. Communication in the presence of noise. Proceedings of the IRE 37, 1 (1949), 10–21

  13. [21]

    David Silver, Julian Schrittwieser, Karen Simonyan, Ioannis Antonoglou, Aja Huang, Arthur Guez, Thomas Hubert, Lucas Baker, Matthew Lai, Adrian Bolton, et al. 2017. Mastering the game of go without human knowledge. nature 550, 7676 (2017), 354–359

  14. [22]

    Geng Sun, Long He, Zemin Sun, Qingqing Wu, Shuang Liang, Jiahui Li, Dusit Niyato, and Victor C. M. Leung. 2024. Joint Task Offloading and Resource Allocation in Aerial-Terrestrial UAV Networks With Edge and Fog Computing for Post-Disaster Rescue.IEEE Transactions on Mobile Com...

  15. [23]

    Hui Sun, Bo Zhang, Xiuye Zhang, Ying Yu, Kewei Sha, and Weisong Shi. 2022. FlexEdge: Dynamic task scheduling for a UAV-based on-demand mobile edge server. IEEE Internet of Things Journal 9, 17 (2022), 15983–16005

  16. [24]

    Hui Sun, Xiuye Zhang, Bo Zhang, Kewei Sha, and Weisong Shi. 2023. Optimal task offloading and trajectory planning algorithms for collaborative video analytics with UAV-assisted edge in disaster rescue. IEEE Transactions on Vehicular Technology (2023)

  17. [25]

    Xiangdong Tang, Fei Chen, Feng Wang, and Zixi Jia. 2023. Disaster Resilient Emergency Communication With Intelligent Air-Ground Cooperation. IEEE Internet of Things Journal (2023)

  18. [26]

    Dinh-Hieu Tran, Symeon Chatzinotas, and Björn Ottersten. 2022. Throughput maximization for backscatter-and cache-assisted wireless powered UAV technology. IEEE Transactions on Vehicular Technology 71, 5 (2022), 5187–5202

  19. [27]

    Shaohua Wan, Songtao Ding, and Chen Chen. 2022. Edge computing enabled video segmentation for real-time traffic monitoring in internet of vehicles. Pattern Recognition 121 (2022), 108146

  20. [28]

    Liang Wang, Kezhi Wang, Cunhua Pan, Wei Xu, Nauman Aslam, and Arumugam Nallanathan. 2021. Deep reinforcement learning based dynamic trajectory control for UAV-assisted mobile edge computing. IEEE Transactions on Mobile Computing 21, 10 (2021), 3536–3550

  21. [29]

    Jianwen Xu, Kaoru Ota, and Mianxiong Dong. 2020. Big data on the fly: UAV-mounted mobile edge computing for disaster management. IEEE Transactions on Network Science and Engineering 7, 4 (2020), 2620–2630

  22. [30]

    Yong Zeng, Rui Zhang, and Teng Joon Lim. 2016. Throughput maximization for UAV-enabled mobile relaying systems.IEEE Transactions on communications 64, 12 (2016), 4983–4996

  23. [31]

    Jun Zhang and Khaled B Letaief. 2019. Mobile edge intelligence and computing for the internet of vehicles. Proc. IEEE 108, 2 (2019), 246–261

  24. [32]

    Liang Zhang, Bijan Jabbari, and Nirwan Ansari. 2022. Deep reinforcement learning driven UAV-assisted edge computing. IEEE Internet of Things Journal 9, 24 (2022), 25449–25459

  25. [33]

    Lu Zhang, Zi-Yan Zhang, Luo Min, Chao Tang, Hong-Ying Zhang, Ya-Hong Wang, and Peng Cai. 2021. Task offloading and trajectory control for UAV-assisted mobile edge computing using deep reinforcement learning. IEEE Access 9 (2021), 53708–53719

  26. [34]

    Nan Zhao, Zhiyang Ye, Yiyang Pei, Ying-Chang Liang, and Dusit Niyato. 2022. Multi-agent deep reinforcement learning for task offloading in UAV-assisted mobile edge computing. IEEE Transactions on Wireless Communications 21, 9 (2022), 6949–6960. ACM J. Auton. Transport. Syst., ...

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.