pith. sign in

arxiv: 2603.07456 · v2 · submitted 2026-03-08 · 💻 cs.DC

Agentic AI-Driven UAV Network Deployment: An LLM-Enhanced Exact Potential Game Approach

Xin Tang , Xiaohuan Li , Qian Chen , Binhan Liao , Yaqi Zhang , Jianxin Chen , Changyuan Zhao , Junchuan Fan
show 1 more author
Junxi Tian
This is my paper

Pith reviewed 2026-05-15 15:25 UTC · model grok-4.3

classification 💻 cs.DC
keywords UAV network deploymentexact potential gamesagentic AILLM utility weightslink configuration optimizationenergy and latencymulti-agent systems
0
0 comments X

The pith

Exact potential games split UAV deployment into discrete links and continuous parameters, with an LLM tuning weights for better energy, latency, and throughput.

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

The paper shows that UAV network deployment, a mixed-integer nonconvex problem from the tight coupling of discrete inter-UAV links and continuous placement/power choices, can be decomposed into two independent exact potential games. At large scale a log-linear learning algorithm finds sparse yet connected link topologies; at small scale an approximate-gradient algorithm jointly tunes UAV positions, transmit powers, and ground-user associations. An LLM then supplies scenario-specific utility weights so the framework adapts across heterogeneous conditions without manual retuning. If the decomposition holds, the resulting multi-agent system reaches stable equilibria faster than centralized or heuristic baselines while lowering energy use and end-to-end delay.

Core claim

A dual-scale exact-potential-game framework solves the UAVN deployment problem by separating large-scale discrete link configuration (via log-linear learning EPG) from small-scale continuous optimization of positions, powers, and associations (via approximate-gradient EPG); an LLM automatically sets the utility-function weights to match network characteristics, yielding lower energy consumption, reduced latency, and higher throughput than baseline methods.

What carries the argument

Exact potential games (EPGs) operating at two spatial scales, with the LLM serving as a knowledge-driven utility-weight generator.

If this is right

  • Link configurations become sparser and less interfering while remaining connected.
  • Joint optimization of placement, power, and association improves throughput and cuts latency.
  • Automatic weight generation removes the need for scenario-specific manual tuning.
  • The multi-agent view treats UAVs as coordinated intelligent agents rather than fixed infrastructure.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same two-scale EPG split could be tested on other coupled discrete-continuous problems such as satellite constellation reconfiguration.
  • If LLM weight generation proves robust, the method might reduce the engineering effort required to deploy UAV networks in new environments.
  • Convergence guarantees of the log-linear and approximate-gradient EPGs could be checked against real flight-test traces to confirm the assumed potential-function property.

Load-bearing premise

The mixed-integer UAV deployment problem can be cleanly split into independent large-scale link and small-scale continuous subproblems via exact potential games, and an LLM can generate unbiased, stable utility weights for any scenario.

What would settle it

A dynamic scenario in which the EPG algorithms fail to converge to a Nash equilibrium or the LLM-generated weights produce higher energy or latency than a hand-tuned baseline.

Figures

Figures reproduced from arXiv: 2603.07456 by Binhan Liao, Changyuan Zhao, Jianxin Chen, Junchuan Fan, Junxi Tian, Qian Chen, Xiaohuan Li, Xin Tang, Yaqi Zhang.

Figure 1
Figure 1. Figure 1: Agentic AI–driven UAV network deployment model. The model [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: The decomposition process of the Agentic AI-driven UAVN deployment problem. Agentic AI is capable of decoupling high-level tasks, generate sub [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Framework of RAG-based Large Language Model assisted UAV network optimization. The framework consists of four modules: [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Retrieval precision under different knowledge block sizes. [PITH_FULL_IMAGE:figures/full_fig_p010_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Verification of convergence and global consistency of algorithms. [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Topology optimization process. (a) 3D topology (b) Node altitude [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Topology of UAV network. slot t + 32). It can be observed that the UAV coordinates have undergone adaptive adjustments, ensuring that all links successfully bypass or fly over obstacles [PITH_FULL_IMAGE:figures/full_fig_p011_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Comparison of network performance. conditions. Compared with the four benchmark algorithms under different network scales, the network throughput is improved by approximately 8.4%, as shown in [PITH_FULL_IMAGE:figures/full_fig_p012_8.png] view at source ↗
read the original abstract

Unmanned aerial vehicular network (UAVN) is envisioned to provide flexible connectivity, wide-area coverage, and low-latency services in dynamic environments. From an agentic artificial intelligence (Agentic AI) perspective, UAVNs naturally operate as multi-agent systems, where UAVs act as intelligent agents that coordinate deployment and networking decisions to achieve global performance objectives. However, the strong coupling between discrete link decisions and continuous deployment parameters makes UAVN deployment optimization a mixed-integer nonconvex problem, resulting in challenges in scalability, efficiency, and solution consistency under dynamic network conditions. This paper proposes a dual spatial-scale UAVN deployment optimization framework based on exact potential games (EPGs), enhanced by Agentic AI. At the large spatial scale, a log-linear learning based EPG (L3-EPG) algorithm is developed to optimize inter-UAV link configurations, enabling sparse yet connected network topologies while reducing redundant links and interference. At the small spatial scale, an approximate gradient based EPG (AG-EPG) algorithm jointly optimizes UAV deployment, transmission power allocation, and ground user (GU) association to improve network throughput and latency. To further enhance adaptability across heterogeneous scenarios, a large language model (LLM) is incorporated as a knowledge-driven decision enhancer to automatically generate utility weights according to network characteristics, alleviating reliance on manual parameter tuning. Simulation results demonstrate that the proposed framework consistently outperforms baseline methods in terms of energy consumption, end-to-end latency, and system throughput.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 1 minor

Summary. The paper proposes a dual spatial-scale UAVN deployment optimization framework using exact potential games (EPGs) enhanced by Agentic AI. At large scale, a log-linear learning EPG (L3-EPG) optimizes inter-UAV link configurations for sparse connected topologies; at small scale, an approximate gradient EPG (AG-EPG) jointly optimizes UAV positions, power allocation, and ground user association. An LLM generates scenario-specific utility weights to improve adaptability without manual tuning. Simulations are reported to show consistent outperformance over baselines in energy consumption, end-to-end latency, and throughput.

Significance. If the simulation results and LLM integration prove robust, the work could advance scalable optimization for dynamic UAV networks by combining game-theoretic decompositions with knowledge-driven AI, addressing mixed-integer nonconvex challenges in multi-agent coordination. The dual-scale EPG approach and LLM weight generation represent a potentially useful direction for reducing reliance on hand-tuned parameters in wireless deployment problems.

major comments (3)
  1. [Abstract] Abstract: The central claim of consistent outperformance in energy, latency, and throughput is presented without any simulation details, baseline definitions, statistical tests, error bars, or scenario descriptions, making it impossible to assess whether gains are robust or scenario-specific.
  2. [LLM-enhanced EPG section] LLM-enhanced EPG section: Utility weights are generated by the LLM according to network characteristics, but no ablation studies, variance analysis across prompts/temperatures/models, or fixed-weight baseline comparisons are described; this leaves open the possibility that reported equilibria and performance improvements are sensitive to LLM outputs rather than the EPG decomposition itself.
  3. [Simulation results] Simulation results: The decoupling into independent large-scale link configuration and small-scale continuous optimization subproblems via EPGs is load-bearing for the scalability claim, yet no verification of potential function consistency or equilibrium stability under LLM weight variation is provided.
minor comments (1)
  1. [Proposed Framework] Notation for L3-EPG and AG-EPG could be clarified with explicit definitions of the potential functions and learning rules in the main text rather than relying solely on references.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed comments. These have highlighted important areas for improving the clarity, robustness, and verifiability of our claims. We address each major comment below and outline the corresponding revisions.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central claim of consistent outperformance in energy, latency, and throughput is presented without any simulation details, baseline definitions, statistical tests, error bars, or scenario descriptions, making it impossible to assess whether gains are robust or scenario-specific.

    Authors: We agree that the abstract would benefit from additional specifics to support the performance claims. In the revised version, we will expand the abstract to briefly describe the simulation scenarios (e.g., UAV counts from 10 to 50, dynamic GU distributions), the baseline methods (random deployment, greedy link selection, and centralized non-game-theoretic optimization), key quantitative gains (e.g., 15-25% energy reduction, 20-35% latency improvement), and note that all results are averaged over 100 Monte Carlo runs with error bars indicating standard deviation. This will make the central claims more transparent and assessable. revision: yes

  2. Referee: [LLM-enhanced EPG section] LLM-enhanced EPG section: Utility weights are generated by the LLM according to network characteristics, but no ablation studies, variance analysis across prompts/temperatures/models, or fixed-weight baseline comparisons are described; this leaves open the possibility that reported equilibria and performance improvements are sensitive to LLM outputs rather than the EPG decomposition itself.

    Authors: We acknowledge that demonstrating robustness to LLM variations is essential. Although the manuscript emphasizes the EPG framework's core benefits, we will add a dedicated ablation subsection in the revised paper. This will include comparisons of LLM-generated weights against fixed-weight and manually tuned baselines, as well as variance analysis across prompt variations, temperatures (0.1-0.9), and alternative models. These additions will confirm that performance gains stem primarily from the EPG decomposition rather than specific LLM outputs. revision: yes

  3. Referee: [Simulation results] Simulation results: The decoupling into independent large-scale link configuration and small-scale continuous optimization subproblems via EPGs is load-bearing for the scalability claim, yet no verification of potential function consistency or equilibrium stability under LLM weight variation is provided.

    Authors: The manuscript already derives the exact potential functions for both L3-EPG and AG-EPG, which mathematically guarantee that unilateral improvements increase the potential and that equilibria are consistent with the global objective, thereby justifying the decoupling. To directly address stability under LLM-induced weight variations, we will include new simulation results in the revised manuscript. These will show convergence trajectories, potential function values, and equilibrium performance metrics when weights are varied according to LLM outputs, confirming that the scalability properties hold. revision: partial

Circularity Check

0 steps flagged

No significant circularity in the derivation chain

full rationale

The paper proposes a dual-scale EPG framework (L3-EPG for link config, AG-EPG for deployment/power/association) with LLM-generated utility weights as an adaptability enhancer. Performance claims rest on simulation comparisons to baselines for energy, latency, and throughput. No equations, definitions, or steps in the abstract or description reduce the claimed results to inputs by construction (e.g., no self-definitional utility weights, no fitted parameters relabeled as predictions, no load-bearing self-citations). The LLM step is presented as external knowledge injection rather than a closed loop; simulation validation remains independent. This is the expected non-finding for a simulation-driven systems paper.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the domain assumption that UAV interactions admit an exact potential game structure allowing clean scale separation, plus the premise that LLM outputs can serve as stable, unbiased utility weights. No explicit free parameters are named, but LLM prompting choices function as implicit tunable elements. No new entities are postulated.

free parameters (1)
  • LLM-generated utility weights
    Automatically produced by the language model according to network characteristics; their specific values are not derived from first principles and directly influence the optimization objective.
axioms (1)
  • domain assumption UAV network deployment can be modeled as a multi-agent system whose discrete link and continuous deployment decisions admit an exact potential game decomposition at two spatial scales.
    Invoked to justify the L3-EPG and AG-EPG split; no proof or reference to prior validation is provided in the abstract.

pith-pipeline@v0.9.0 · 5592 in / 1353 out tokens · 45711 ms · 2026-05-15T15:25:43.204906+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

43 extracted references · 43 canonical work pages · 1 internal anchor

  1. [1]

    Toward edge general intelligence with agentic ai and agentification: Concepts, technologies, and future directions,

    R. Zhang, G. Liu, Y . Liu, C. Zhao, J. Wang, Y . Xu, D. Niyato, J. Kang, Y . Li, S. Maoet al., “Toward edge general intelligence with agentic ai and agentification: Concepts, technologies, and future directions,”IEEE Communications Surveys & Tutorials, vol. 28, pp. 4285–4318, 2026

    work page 2026

  2. [2]

    Multi-objective optimization for uav swarm-assisted iot with virtual antenna arrays,

    J. Li, G. Sun, L. Duan, and Q. Wu, “Multi-objective optimization for uav swarm-assisted iot with virtual antenna arrays,”IEEE Transactions on Mobile Computing, vol. 23, no. 5, pp. 4890–4907, 2023

    work page 2023

  3. [3]

    Digital twin-empowered task assignment in aerial mec network: A resource coalition cooperation ap- proach with generative model,

    X. Tang, Q. Chen, R. Yu, and X. Li, “Digital twin-empowered task assignment in aerial mec network: A resource coalition cooperation ap- proach with generative model,”IEEE Transactions on Network Science and Engineering, 2024

    work page 2024

  4. [4]

    Interactive ai with retrieval-augmented generation for next generation networking,

    R. Zhang, H. Du, Y . Liu, D. Niyato, J. Kang, S. Sun, X. Shen, and H. V . Poor, “Interactive ai with retrieval-augmented generation for next generation networking,”IEEE Network, vol. 38, no. 6, pp. 414–424, 2024

    work page 2024

  5. [5]

    Tradeoff between age of information and operation time for uav sensing over multi-cell cellular networks,

    C. Zhan, H. Hu, J. Wang, Z. Liu, and S. Mao, “Tradeoff between age of information and operation time for uav sensing over multi-cell cellular networks,”IEEE Transactions on Mobile Computing, vol. 23, no. 4, pp. 2976–2991, 2024

    work page 2024

  6. [6]

    Joint task offloading and resource allocation in aerial-terrestrial uav networks with edge and fog computing for post-disaster rescue,

    G. Sun, L. He, Z. Sun, Q. Wu, S. Liang, J. Li, D. Niyato, and V . C. Leung, “Joint task offloading and resource allocation in aerial-terrestrial uav networks with edge and fog computing for post-disaster rescue,” IEEE Transactions on Mobile Computing, vol. 23, no. 9, pp. 8582–8600, 2024

    work page 2024

  7. [7]

    Energy- efficient communication networks via multiple aerial reconfigurable intelligent surfaces: Drl and optimization approach,

    J. Wang, L. Bai, Z. Fang, R. Han, J. Wang, and J. Choi, “Energy- efficient communication networks via multiple aerial reconfigurable intelligent surfaces: Drl and optimization approach,”IEEE Transactions on Vehicular Technology, vol. 73, no. 7, pp. 9755–9770, 2024

    work page 2024

  8. [9]

    Joint optimization of deployment and flight planning of multi-uavs for long- distance data collection from large-scale iot devices,

    Y . Zhang, Y . Huang, C. Huang, H. Huang, and A.-T. Nguyen, “Joint optimization of deployment and flight planning of multi-uavs for long- distance data collection from large-scale iot devices,”IEEE Internet of Things Journal, vol. 11, no. 1, pp. 791–804, 2024

    work page 2024

  9. [10]

    Location privacy-aware and energy- efficient offloading for distributed edge computing,

    Y . He, X. He, R. Jin, and H. Dai, “Location privacy-aware and energy- efficient offloading for distributed edge computing,”IEEE Transactions on Wireless Communications, vol. 22, no. 11, pp. 7975–7988, 2023

    work page 2023

  10. [11]

    Joint trajectory and power optimization for leo satellite-assisted uav communications: A stackelberg game approach,

    Y . Gao, J. Tang, B. Li, D. Niyato, and D. I. Kim, “Joint trajectory and power optimization for leo satellite-assisted uav communications: A stackelberg game approach,”IEEE Transactions on Communications, vol. 72, no. 1, pp. 538–552, 2024

    work page 2024

  11. [12]

    A novel hybrid enhanced particle swarm optimization for uav path planning,

    J. Wu, Y . Sun, J. Bi, and C. Chen, “A novel hybrid enhanced particle swarm optimization for uav path planning,”IEEE Transactions on Vehicular Technology, 2025

    work page 2025

  12. [13]

    Hybrid heuristic-based multi-uav route planning for time-dependent data collection,

    P. Wan, S. Wang, G. Xu, Y . Long, and R. Hu, “Hybrid heuristic-based multi-uav route planning for time-dependent data collection,”IEEE Internet of Things Journal, vol. 11, no. 13, pp. 24 134–24 147, 2024

    work page 2024

  13. [14]

    Bi-objective ant colony optimization for trajectory planning and task offloading in uav-assisted mec systems,

    Y . Wang, J. Zhu, H. Huang, and F. Xiao, “Bi-objective ant colony optimization for trajectory planning and task offloading in uav-assisted mec systems,”IEEE Transactions on Mobile Computing, vol. 23, no. 12, pp. 12 360–12 377, 2024

    work page 2024

  14. [15]

    Joint power allocation and 3d deployment for uav-bss: A game theory based deep reinforcement learning approach,

    S. Fu, X. Feng, A. Sultana, and L. Zhao, “Joint power allocation and 3d deployment for uav-bss: A game theory based deep reinforcement learning approach,”IEEE Transactions on Wireless Communications, vol. 23, no. 1, pp. 736–748, 2023

    work page 2023

  15. [16]

    Deep reinforcement learning based resource allocation in multi-uav- aided mec networks,

    J. Chen, X. Lyu, K. Wang, Y . Zhang, K. Li, T. Hong, and C. S. Hong, “Deep reinforcement learning based resource allocation in multi-uav- aided mec networks,”IEEE Transactions on Communications, vol. 71, no. 1, pp. 296–309, 2023

    work page 2023

  16. [17]

    Multi-agent deep reinforcement learning for joint decoupled user association and trajectory design in full- duplex multi-uav networks,

    C. Dai, K. Zhu, and E. Hossain, “Multi-agent deep reinforcement learning for joint decoupled user association and trajectory design in full- duplex multi-uav networks,”IEEE Transactions on Mobile Computing, vol. 22, no. 10, pp. 6056–6070, 2023

    work page 2023

  17. [18]

    Bayesian optimization enhanced deep reinforcement learning for tra- jectory planning and network formation in multi-uav networks,

    S. Gong, M. Wang, B. Gu, W. Zhang, D. T. Hoang, and D. Niyato, “Bayesian optimization enhanced deep reinforcement learning for tra- jectory planning and network formation in multi-uav networks,”IEEE Transactions on Vehicular Technology, vol. 72, no. 8, pp. 10 933–10 948, 2023

    work page 2023

  18. [19]

    Formation-aware uav net- work self-organization with game-theoretic distributed topology con- trol,

    Z. Du, S. Wang, X. Wang, and J. Chen, “Formation-aware uav net- work self-organization with game-theoretic distributed topology con- trol,”IEEE Transactions on Cognitive Communications and Networking, vol. 11, no. 5, pp. 3470–3485, 2025

    work page 2025

  19. [20]

    Multi-agent rein- forcement learning based uav swarm communications against jamming,

    Z. Lv, L. Xiao, Y . Du, G. Niu, C. Xing, and W. Xu, “Multi-agent rein- forcement learning based uav swarm communications against jamming,” IEEE Transactions on Wireless Communications, vol. 22, no. 12, pp. 9063–9075, 2023

    work page 2023

  20. [21]

    Uav assisted inte- grated sensing and communications for internet of things: 3d trajectory optimization and resource allocation,

    Z. Liu, X. Liu, Y . Liu, V . C. Leung, and T. S. Durrani, “Uav assisted inte- grated sensing and communications for internet of things: 3d trajectory optimization and resource allocation,”IEEE Transactions on Wireless Communications, vol. 23, no. 8, pp. 9346–9361, 2024

    work page 2024

  21. [22]

    Throughput maximization for uav-enabled integrated IEEE TRANSACTIONS ON XX , VOL. X, NO. X, X 2026 13 periodic sensing and communication,

    K. Meng, Q. Wu, J. Xu, W. Chen, Z. Feng, R. Schober, and A. L. Swindlehurst, “Throughput maximization for uav-enabled integrated IEEE TRANSACTIONS ON XX , VOL. X, NO. X, X 2026 13 periodic sensing and communication,”IEEE Transactions on Wireless Communications, vol. 22, no. 1, pp. 671–687, 2023

    work page 2026

  22. [23]

    Reinforcement learning based energy-efficient fast routing for fanets,

    J. Li, L. Xiao, X. Qi, Z. Lv, Q. Chen, and Y .-J. Liu, “Reinforcement learning based energy-efficient fast routing for fanets,”IEEE Transac- tions on Communications, vol. 72, no. 11, pp. 7063–7076, 2024

    work page 2024

  23. [24]

    Decentralized reinforcement learning in markov potential games,

    R. Zhang, Z. Ren, and N. Li, “Decentralized reinforcement learning in markov potential games,”IEEE Transactions on Automatic Control, vol. 69, no. 11, pp. 7867–7874, 2024

    work page 2024

  24. [25]

    Dynamic topology organization and maintenance algorithms for autonomous uav swarms,

    A. Gaydamaka, A. Samuylov, D. Moltchanov, M. Ashraf, B. Tan, and Y . Koucheryavy, “Dynamic topology organization and maintenance algorithms for autonomous uav swarms,”IEEE Transactions on Mobile Computing, vol. 23, no. 5, pp. 4423–4439, 2024

    work page 2024

  25. [26]

    Optimal measurement policy for linear mea- surement systems with applications to uav network topology prediction,

    Y . Sadi and S. C. Ergen, “Optimal measurement policy for linear mea- surement systems with applications to uav network topology prediction,” IEEE Transactions on Vehicular Technology, vol. 69, no. 2, pp. 1970– 1981, Feb 2020

    work page 1970

  26. [27]

    Exploring bio- logical robustness for reliable multi-uav networks,

    S. K. Yoo, S. L. Cotton, L. Zhang, and W. G. Scanlon, “Exploring bio- logical robustness for reliable multi-uav networks,”IEEE Transactions on Wireless Communications, vol. 20, no. 12, pp. 8034–8047, Dec 2021

    work page 2021

  27. [28]

    Los coverage analysis for uav-based thz communication networks: Toward 3-d visualization of wireless networks,

    M. Polese, M. Giordani, T. Zugno, A. Venugopal, and M. Zorzi, “Los coverage analysis for uav-based thz communication networks: Toward 3-d visualization of wireless networks,”IEEE Transactions on Wireless Communications, vol. 22, no. 12, pp. 8782–8796, Dec 2023

    work page 2023

  28. [29]

    Formation-aware uav net- work self-organization with game-theoretic distributed topology con- trol,

    Z. Du, S. Wang, X. Wang, and J. Chen, “Formation-aware uav net- work self-organization with game-theoretic distributed topology con- trol,”IEEE Transactions on Cognitive Communications and Networking, vol. 11, no. 5, pp. 3470–3485, Oct 2025

    work page 2025

  29. [30]

    Resource scheduling for uavs-aided d2d networks: A multi-objective optimization approach,

    H. Pan, Y . Liu, G. Sun, P. Wang, and C. Yuen, “Resource scheduling for uavs-aided d2d networks: A multi-objective optimization approach,” IEEE Transactions on Wireless Communications, vol. 23, no. 5, pp. 4691–4708, 2023

    work page 2023

  30. [31]

    Joint optimization of deployment and flight planning of multi-uavs for long- distance data collection from large-scale iot devices,

    Y . Zhang, Y . Huang, C. Huang, H. Huang, and A.-T. Nguyen, “Joint optimization of deployment and flight planning of multi-uavs for long- distance data collection from large-scale iot devices,”IEEE Internet of Things Journal, vol. 11, no. 1, pp. 791–804, 2023

    work page 2023

  31. [32]

    Enhancing 6g iot communications: Uav relayed satellite d2d with large language model optimization,

    L. Liu, W. Li, and F. Zhou, “Enhancing 6g iot communications: Uav relayed satellite d2d with large language model optimization,”IEEE Communications Magazine, 2025

    work page 2025

  32. [33]

    Generative ai agents with large language model for satellite networks via a mixture of experts transmission,

    R. Zhang, H. Du, Y . Liu, D. Niyato, J. Kang, Z. Xiong, A. Jamalipour, and D. I. Kim, “Generative ai agents with large language model for satellite networks via a mixture of experts transmission,”IEEE Journal on Selected Areas in Communications, 2024

    work page 2024

  33. [34]

    Optimizing 6g integrated sensing and communications (isac) via expert networks,

    J. Wang, H. Du, G. Sun, J. Kang, H. Zhou, D. Niyato, and J. Chen, “Optimizing 6g integrated sensing and communications (isac) via expert networks,”arXiv preprint arXiv:2406.00408, 2024

    work page arXiv 2024

  34. [35]

    Ieee standard for a framework for structuring low-altitude airspace for unmanned aerial vehicle (uav) operations,

    IEEE Communication Society, “Ieee standard for a framework for structuring low-altitude airspace for unmanned aerial vehicle (uav) operations,” IEEE, Tech. Rep. IEEE Std 1939.1-2021, 2021

    work page 1939

  35. [36]

    Propagation data and prediction methods required for the design of terrestrial broadband radio access systems operating in a frequency range from 3 to 60 ghz,

    ITU-R, “Propagation data and prediction methods required for the design of terrestrial broadband radio access systems operating in a frequency range from 3 to 60 ghz,” International Telecommunication Union, Tech. Rep. Recommendation ITU-R P.1410-5, 2019

    work page 2019

  36. [37]

    Multi-agent deep reinforcement learning for task offloading in uav-assisted mobile edge computing,

    Z. Zhao, Y . Feng, J. Li, Y . Xu, and N. Cheng, “Multi-agent deep reinforcement learning for task offloading in uav-assisted mobile edge computing,”IEEE Transactions on Wireless Communications, vol. 21, no. 9, pp. 6949–6960, 2022

    work page 2022

  37. [38]

    Coalition formation game for joint spectrum sensing and channel access in cognitive radio networks,

    Y . Chen, Y . Wang, J. Zhang, and L. Song, “Coalition formation game for joint spectrum sensing and channel access in cognitive radio networks,” IEEE Transactions on Wireless Communications, vol. 20, no. 2, pp. 1496–1510, 2021

    work page 2021

  38. [39]

    New and improved embedding model,

    OpenAI, “New and improved embedding model,” https://openai.com/ blog/new-and-improved-embedding-model, 2022

    work page 2022

  39. [40]

    Chat gpt & google bard ai: A review,

    S. K. Singh, S. Kumar, and P. S. Mehra, “Chat gpt & google bard ai: A review,” in2023 International Conference on IoT, Communication and Automation Technology (ICICAT). IEEE, 2023, pp. 1–6

    work page 2023

  40. [41]

    Grok: A large language model,

    xAI, “Grok: A large language model,” https://x.ai/blog/grok, 2024

    work page 2024

  41. [42]

    DeepSeek LLM: Scaling Open-Source Language Models with Longtermism

    DeepSeek-AI, “Deepseek llm: Scaling open-source language models with longtermism,”arXiv preprint arXiv:2401.02954, 2024

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  42. [43]

    Evolutionary games in wireless networks,

    H. Tembine, E. Altman, R. ElAzouzi, and Y . Hayel, “Evolutionary games in wireless networks,”IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), vol. 40, no. 3, pp. 634–646, 2010

    work page 2010

  43. [44]

    Multi-uav- enabled load-balance mobile-edge computing for iot networks,

    Y . Yang, J. Wu, L. Huang, X. Xu, Q. Wu, and R. Schober, “Multi-uav- enabled load-balance mobile-edge computing for iot networks,”IEEE Internet of Things Journal, vol. 10, no. 16, pp. 14 463–14 478, 2023

    work page 2023