Enhanced multi-agent deep reinforcement learning for eﬃcient task oﬄoading and resource allocation in vehicular networks✩

Long Xu; Jiale Tan; Hongcheng Zhuang

doi:10.1016/j.dcan.2025.09.002

›› 2026, Vol. 12 ›› Issue (1) :66 -75. DOI: 10.1016/j.dcan.2025.09.002

Regular Papers

research-article

Enhanced multi-agent deep reinforcement learning for eﬃcient task oﬄoading and resource allocation in vehicular networks^✩

Long Xu ^a
, Jiale Tan ^a
, Hongcheng Zhuang ^a^,^b^,^*^,¹

Author information +

^a School of Electronics and Communication Engineering, Sun Yat-sen University, Shenzhen 518107, China

^b Peng Cheng Laboratory, Shenzhen 518055, Guangdong, China

^* School of Electronics and Communication Engineering, Sun Yat-sen University, Shenzhen 518107, China. E-mail addresses: xulong3@mail2.sysu.edu.cn (L. Xu), zhuanghch@mail.sysu.edu.cn (H. Zhuang).

Long Xu received the B.S. degree in Communication Engineering from Sun Yat-sen University, Shenzhen, China, in 2022. He is currently pursuing the Ph.D. degree in the School of Electronics and Communication Engineering, Sun Yat-sen University, Shenzhen, China. His research interests focus on edge computing, machine learning, and reinforce-ment learning.

Jiale Tan received the B.S. degree in Communication Engineering from Sun Yat-sen University, Shenzhen, China, in 2024. She is currently pursuing the master’s degree in the School of Electronics and Communication Engineering, Sun Yat-sen University, Shenzhen, China. Her research interests include mobile edge computing, computation oﬄoading, and satellite communication.

Hongcheng Zhuang is currently a professor with the School of Electronics and Com-munication Engineering, Sun Yat-sen University, Shenzhen, China. He received the Ph.D. degree in radiophysics from Sun Yat-sen University in 2002. In 2002, he joined Huawei Technologies and led standard research projects, long-term research projects, and major programs in China in wireless communications. He holds more than 300 granted patents in wireless communication. His current research interests include mobile communication system design and advanced wireless technologies, such as intelligent radio and reconfig-urable intelligent surfaces.

Show less

History +

PDF

Abstract

In response to the rising demand for low-latency, computation-intensive applications in vehicular networks, this paper proposes an adaptive task oﬄoading approach for Vehicle-to-Everything (V2X) environments. Leveraging an enhanced Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithm with an attention mechanism, the proposed approach optimizes computation oﬄoading and resource allocation, aiming to minimize energy consumption and service delay. In this paper, vehicles dynamically oﬄoad computing-intensive tasks to both nearby vehicles through V2V links and roadside units through V2I links. The adaptive attention mechanism enables the system to prioritize relevant state information, leading to faster convergence. Simulations conducted in a realistic urban V2X scenario demonstrate that the proposed Attention-enhanced MADDPG (AT-MADDPG) algorithm significantly improves performance, achieving notable reductions in both energy consumption and latency compared to baseline algorithms, especially in high-demand, dynamic scenarios.

Keywords

Computation oﬄoading / Vehicular networks / Deep reinforcement learning / Adaptive oﬄoading / Spectrum and power allocation

Cite this article

Download citation ▾

Long Xu, Jiale Tan, Hongcheng Zhuang. Enhanced multi-agent deep reinforcement learning for eﬃcient task oﬄoading and resource allocation in vehicular networks^✩. , 2026, 12(1): 66-75 DOI:10.1016/j.dcan.2025.09.002

登录浏览全文

4963

注册一个新账户忘记密码

CRediT authorship contribution statement

Long Xu: Writing-original draft, Validation, Investigation, Conceptualization. Jiale Tan: Formal analysis. Hongcheng Zhuang: Writing-review & editing, Supervision, Methodology, Conceptualization.

Declaration of competing interest

This paper, “Enhanced multi-agent deep reinforcement learning for efficient task offloading and resource allocation in vehicular networks“, has no conflict interest.

Acknowledgement

This work was supported by the National Key Research and Development Program of China under grant 2021YFA0716600. (Corresponding author: Hongcheng Zhuang)

References

Publishing order | Descend order by publishing year | Descend order by cited within

[1]	A. Haydari, Y. Yılmaz, Deep reinforcement learning for intelligent transportation systems: a survey, IEEE Trans. Intell. Transp. Syst. 23 (1) (2022) 11-32.

[2]	M. Noor-A-Rahim, Z. Liu, H. Lee, M.O. Khyam, J. He, D. Pesch, K. Moessner, W. Saad, H.V. Poor, 6g for vehicle-to-everything (v2x) communications: enabling tech-nologies, challenges, and opportunities, Proc. IEEE 110 (6) (2022) 712-734.

[3]	M.A. Mirza, Y. Junsheng, S. Raza, M. Ahmed, M. Asif, A. Irshad, N. Kumar, Mcla task oﬄoading framework for 5g-nr-v2x-based heterogeneous vecns, IEEE Trans. Intell. Transp. Syst. 24 (12) (2023) 14329-14346.

[4]	P. Mach, Z. Becvar, Mobile edge computing: a survey on architecture and computa-tion oﬄoading, IEEE Commun. Surv. Tutor. 19 (3) (2017) 1628-1656.

[5]	M. Ahmed, S. Raza, M.A. Mirza, A. Aziz, M.A. Khan, W.U. Khan, J. Li, Z. Han, A survey on vehicular task oﬄoading: classification, issues, and challenges, J. King Saud Univ, Comput. Inf. Sci. 34 (7) (2022) 4135-4162.

[6]	K.B. Letaief, Y. Shi, J. Lu, J. Lu, Edge artificial intelligence for 6g: vision, enabling technologies, and applications, IEEE J. Sel. Areas Commun. 40 (1) (2022) 5-36.

[7]	L. Liang, H. Ye, G.Y. Li, Spectrum sharing in vehicular networks based on multi-agent reinforcement learning, IEEE J. Sel. Areas Commun. 37 (10) (2019) 2282-2292.

[8]	Y. Qi, Y. Zhou, Y.-F. Liu, L. Liu, Z. Pan, Traﬃc-aware task oﬄoading based on con-vergence of communication and sensing in vehicular edge computing, IEEE Internet Things J. 8 (24) (2021) 17762-17777.

[9]	Q. Wu, S. Wang, H. Ge, P. Fan, Q. Fan, K.B. Letaief, Delay-sensitive task oﬄoading in vehicular fog computing-assisted platoons, IEEE Trans. Netw. Serv. Manag. 21 (2) (2023) 2012-2026.

[10]	Y. Ju, Y. Chen, Z. Cao, L. Liu, Q. Pei, M. Xiao, K. Ota, M. Dong, V.C.M. Leung, Joint secure oﬄoading and resource allocation for vehicular edge computing network: a multi-agent deep reinforcement learning approach, IEEE Trans. Intell. Transp. Syst. 24 (5) (2023) 5555-5569.

[11]	S.A. Ullah, M. Bibi, S.A. Hassan, H. Abou-Zeid, H.K. Qureshi, H. Jung, A. Mah- mood, M. Gidlund, E. Hossain, From nodes to roads: surveying drl applications in mec-enhanced terrestrial wireless networks, IEEE Commun. Surv. Tutor. 28 (2026) 1169-1208.

[12]	Q. He, G. Cui, X. Zhang, F. Chen, S. Deng, H. Jin, Y. Li, Y. Yang, A game-theoretical approach for user allocation in edge computing environment, IEEE Trans. Parallel Distrib. Syst. 31 (3) (2019) 515-529.

[13]	M.K. Hasan, N. Jahan, M.Z.A. Nazri, S. Islam, M.A. Khan, A.I. Alzahrani, N. Alalwan, Y. Nam, Federated learning for computational oﬄoading and resource management of vehicular edge computing in 6g-v2x network, IEEE Trans. Consum. Electron. 70 (1) (2024) 3827-3847.

[14]	X. Li, L. Ma, R. Shankaran, Y. Xu, M.A. Orgun, Joint power control and resource allocation mode selection for safety-related v2x communication, IEEE Trans. Veh. Technol. 68 (8) (2019) 7970-7986.

[15]	H. Zhang, X. Liu, Y. Xu, D. Li, C. Yuen, Q. Xue, Partial oﬄoading and resource allocation for mec-assisted vehicular networks, IEEE Trans. Veh. Technol. 73 (1) (2024) 1276-1288.

[16]	H. Materwala, L. Ismail, R.M. Shubair, R. Buyya, Energy-sla-aware genetic algorithm for edge-cloud integrated computation oﬄoading in vehicular networks, Future Gener. Comput. Syst. 135 (2022) 205-222.

[17]	H. Materwala, L. Ismail, H.S. Hassanein, Qos-sla-aware adaptive genetic algorithm for multi-request oﬄoading in integrated edge-cloud computing in internet of vehi-cles, Veh. Commun. 43 (2023) 100654.

[18]	Y. Zhang, C. Chen, H. Zhu, Y. Pan, J. Wang, Latency minimization for mec-v2x as-sisted autonomous vehicles task oﬄoading, IEEE Trans. Veh. Technol. 74 (3) (2025) 4917-4932.

[19]	H. Zhou, Z. Wang, N. Cheng, D. Zeng, P. Fan, Stackelberg-game-based computation oﬄoading method in cloud-edge computing networks, IEEE Internet Things J. 9 (17) (2022) 16510-16520.

[20]	D. Zhang, L. Tan, J. Ren, M.K. Awad, S. Zhang, Y. Zhang, P.-J. Wan, Near-optimal and truthful online auction for computation oﬄoading in green edge-computing systems, IEEE Trans. Mob. Comput. 19 (4) (2020) 880-893.

[21]	Y. Liang, H. Tang, H. Wu, Y. Wang, P. Jiao, Lyapunov-guided oﬄoading optimiza-tion based on soft actor-critic for Isac-aided internet of vehicles, IEEE Trans. Mob. Comput. 23 (12) (2024) 14708-14721.

[22]	J. Wang, C. Jiang, H. Zhang, Y. Ren, K.-C. Chen, L. Hanzo, Thirty years of machine learning: the road to Pareto-optimal wireless networks, IEEE Commun. Surv. Tutor. 22 (3) (2020) 1472-1514.

[23]	Y. Ju, Y. Chen, Z. Cao, L. Liu, Q. Pei, M. Xiao, K. Ota, M. Dong, V.C. Leung, Joint secure oﬄoading and resource allocation for vehicular edge computing network: a multi-agent deep reinforcement learning approach, IEEE Trans. Intell. Transp. Syst. 24 (5) (2023) 5555-5569.

[24]	W. Feng, S. Lin, N. Zhang, G. Wang, B. Ai, L. Cai, Joint c-v2x based oﬄoading and resource allocation in multi-tier vehicular edge computing system, IEEE J. Sel. Areas Commun. 41 (2) (2023) 432-445.

[25]	X. Song, Q. Chen, S. Wang, T. Song, Cross-domain resources optimization for hybrid edge computing, Digit. Commun. Netw. 11 (6) (2025) 1797-1808.

[26]	M. Li, J. Gao, N. Zhang, L. Zhao, X. Shen,Collaborative computing in vehicular networks: a deep reinforcement learning approach, in: ICC 2020-2020 IEEE Inter-national Conference on Communications (ICC), 2020, pp. 1-6.

[27]	J. Du, L. Wang, Y. Lin, P. Qian, Vehicle-road cooperative task oﬄoading with task migration in mec-enabled iov, in:International Conference on Wireless Algorithms, Systems, and Applications, Springer, 2022, pp. 261-272.

[28]	M. Ahmed, J. Liu, M.A. Mirza, W.U. Khan, F.N. Al-Wesabi, Marl based resource al-location scheme leveraging vehicular cloudlet in automotive-industry 5.0, J. King Saud Univ, Comput. Inf. Sci. 35 (6) (2023) 101420.

[29]	Z. Yao, S. Xia, Y. Li, G. Wu, Cooperative task oﬄoading and service caching for digital twin edge networks: a graph attention multi-agent reinforcement learning approach, IEEE J. Sel. Areas Commun. 41 (11) (2023) 3401-3413.

[30]	B. Hu, W. Zhang, Y. Gao, J. Du, X. Chu, Multi-agent deep deterministic policy gradient-based computation oﬄoading and resource allocation for Isac-aided 6g v2x networks, IEEE Internet Things J. (2024) 1.

[31]	Z. Chen, L. Zhang, Y. Pei, C. Jiang, L. Yin, Noma-based multi-user mobile edge com-putation oﬄoading via cooperative multi-agent deep reinforcement learning, IEEE Trans. Cogn. Commun. Netw. 8 (1) (2021) 350-364.