SAC-Based Collaborative Task Offloading and Path Optimization for Vehicular Edge Computing

Wu Wang

School of Mathematics and Computer Science, Yunnan Minzu University, Kunming 650504, China

*Corresponding author

With the rapid development of Vehicle-to-Everything (V2X) and Vehicular Edge Computing (VEC), the massive computation-intensive tasks generated by intelligent vehicles during driving, including environmental perception, path planning, and autonomous driving decision-making, impose extremely high requirements on real-time performance and reliability. Traditional task offloading strategies often consider communication quality or computing resources in isolation, thereby neglecting the inherent coupling between a vehicle’s dynamic driving path and its offloading decisions. This leads to low task completion rates, long travel times, and insufficient utilization of edge computing resources. Therefore, addressing the joint optimization problem in VEC scenarios, this paper proposes a deep reinforcement learning (DRL) algorithm based on an improved hybrid action space Soft Actor-Critic (SAC-Discrete). An agent capable of simultaneously handling continuous speed control and discrete offloading decisions is designed, enabling it to learn to collaboratively trade off task offloading and path optimization in dynamic traffic environments. This paper systematically compares the proposed algorithm with five baseline methods: greedy strategy, fixed path optimization, fixed offloading optimization, random strategy, and a discretized-action version of Vanilla SAC. Experimental results demonstrate that our algorithm converges rapidly and achieves high stability, and the proposed joint optimization framework significantly improves system efficiency and resource utilization while reducing average travel time.

VEC, Task Offloading, SAC-Discrete, DRL, Collaborative Optimization

Jun Wang, Lin Chai, Yumei Yang, and Wu Wang (2026). SAC-Based Collaborative Task Offloading and Path Optimization for Vehicular Edge Computing. Journal of Networking and Network Applications, Volume 6, Issue 1, pp. 20–31. https://doi.org/10.33969/J-NaNA.2026.060103.

[1] J. Shi, J. Du, J. Wang, and J. Yuan, “Deep reinforcement learning-based v2v partial computation offloading in vehicular fog computing,” in 2021 IEEE Wireless Communications and Networking Conference (WCNC), 2021, pp. 1–6.

[2] Q. Wu, H. Liu, R. Wang, P. Fan, Q. Fan, and Z. Li, “Delay-sensitive task offloading in the 802.11p-based vehicular fog computing systems,” IEEE Internet of Things Journal, vol. 7, no. 1, pp. 773–785, 2020.

[3] X. Hou, Y. Li, M. Chen, D. Wu, D. Jin, and S. Chen, “Vehicular fog computing: A viewpoint of vehicles as the infrastructures,” IEEE Transactions on Vehicular Technology, vol. 65, no. 6, pp. 3860–3873, 2016.

[4] W. Saad, M. Bennis, and M. Chen, “A vision of 6g wireless systems: Applications, trends, technologies, and open research problems,” IEEE Network, vol. 34, no. 3, pp. 134–142, 2020.

[5] K. Zhang, M. Peng, and Y. Sun, “Delay-optimized resource allocation in fog-based vehicular networks,” IEEE Internet of Things Journal, vol. 8, no. 3, pp. 1347–1357, 2021.

[6] S. Sharma, S. Ghosh, M. R. Bhatnagar, and B. K. Panigrahi, “Game-theoretic flexible resource allocation for handoff in hybrid v2x com-munication,” in 2023 IEEE 34th Annual International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC), 2023, pp. 1–7.

[7] H. Park, Y. Jang, and J.-W. Choi, “Latency analysis of 5g c-v2x real-time video transmission over different channel states,” in 2025 IEEE 101st Vehicular Technology Conference (VTC2025-Spring), 2025, pp. 1–6.

[8] Y. Eunju, K. Bongseob, S. Geunhan, Y. Kyungsu, and H. Sujin, “V2x traffic-load generator system for enabling efficient vehicle commu-nication load-balancing,” in 2023 14th International Conference on Information and Communication Technology Convergence (ICTC), 2023, pp. 745–749.

[9] I. Lee and D. K. Kim, “Resource allocation in nr-v2x mode 2 using multi agent dqn,” in 2023 Fourteenth International Conference on Ubiquitous and Future Networks (ICUFN), 2023, pp. 17–19.

[10] Z. Mlika and S. Cherkaoui, “Deep deterministic policy gradient to minimize the age of information in cellular v2x communications,” IEEE Transactions on Intelligent Transportation Systems, vol. 23, no. 12, pp. 23 597–23 612, 2022.

[11] S. Yu, S. Li, Y. Li, Y. Fan, Y. Zhou, and J. Peng, “A risk-barrier soft actor-critic approach for automotive intelligent cruise control towards v2x communication,” in 2025 9th CAA International Conference on Vehicular Control and Intelligence (CVCI), 2025, pp. 1–6.

[12] Z. Zhou, P. Liu, J. Feng, Y. Zhang, S. Mumtaz, and J. Rodriguez, “Computation resource allocation and task assignment optimization in vehicular fog computing: A contract-matching approach,” IEEE Trans-actions on Vehicular Technology, vol. 68, no. 4, pp. 3113–3125, 2019.

[13] Annu and P. Rajalakshmi, “Distance-based queue modeling in sidelink c- v2x communication for platooning: Towards 6g-v2x,” in 2024 IEEE 100th Vehicular Technology Conference (VTC2024-Fall), 2024, pp. 1–5.

[14] L. Ganeshkumar, R. Ramachandran, and T. Ohtsuki, “Delay-constrained sum-rate maximization in star-ris v2x communications using lyapunov framework,” IEEE Access, vol. 14, pp. 857–874, 2026.

[15] P. Zhou, X. Xing, X. Wang, T. Wang, and G. Wang, “Adaptive task offloading in high-speed vehicular networks using enhanced d3qn algorithm,” in 2024 IEEE International Symposium on Parallel and Distributed Processing with Applications (ISPA), 2024, pp. 607–614.

[16] X. Bi, J. Shi, B. Zhang, Z. Lyu, and L. Huang, “An rsu-crossed dependent task offloading scheme for vehicular edge computing based on deep reinforcement learning,” International Journal of Sensor Networks, vol. 41, no. 4, pp. 244–256, 2023.

[17] K. Moghaddasi and R. Jurdak, “An energy-aware distributed federated soft actor-critic framework for intelligent task offloading in vehicular mobile edge computing networks,” Ad Hoc Networks, vol. 180, p. 104043, 2026. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S1570870525002914

[18] V. R. de Lima and M. L. de Campos, “Fully distributed multi-agent processing strategy applied to vehicular networks,” Vehicular Communications, vol. 49, p. 100806, 2024. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S2214209624000810

[19] Q. Zhong, N. Ye, and S. Gao, “A hyper-heuristic algorithm for joint task offloading, resource allocation and trajectory design in uav-assisted vehicular networks,” in International Conference on SmartRail, Traffic and Transportation Engineering. Springer, 2024, pp. 188–194.

[20] N. Keshari and D. Singh, “Tcv-d: An approach for path selection in vehicular task offloading,” Vehicular Communications, vol. 47, p. 100770, 2024. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S2214209624000457

[21] C. S.-H. Hsu, J. Mart´ın-P´erez, D. De Vleeschauwer, L. Valcarenghi, X. Li, and C. Papagianni, “A deep rl approach on task placement and scaling of edge resources for cellular vehicle-to-network service provisioning,” IEEE Transactions on Network and Service Management, 2025.

[22] T. Haarnoja, A. Zhou, P. Abbeel, and S. Levine, “Soft actor-critic: Off-policy maximum entropy deep reinforcement learning with a stochastic actor,” in International conference on machine learning. Pmlr, 2018, pp. 1861–1870.

[23] S. Fujimoto, H. Hoof, and D. Meger, “Addressing function approxi-mation error in actor-critic methods,” in International conference on machine learning. PMLR, 2018, pp. 1587–1596.