Article Details

Abstract

Aiming at the problems of large task data volume caused by excessive vehicle service applications, which leads to high latency, high energy consumption, and failure to process tasks in time, this paper proposes a task offloading and resource allocation strategy based on Deep Reinforcement Learning (DRL). Firstly, this paper designs the Internet of Vehicles (IoV) system framework and proposes a task offloading scheme based on cloud-edge-local collaboration. Secondly, it determines the offloading strategy according to task priority, and constructs mathematical models of edge and local task computing in terms of latency and energy consumption. Finally, taking energy consumption and latency as decision objectives, it constructs an objective function to minimize the total user cost. Experimental results show that under the IoV environment simulation experiment, this method performs better than traditional task offloading strategies, significantly reduces system latency, and improves resource utilization rate.

Keywords

IoV; DRL; task offloading; resource allocation

References

1. [1] Feng Hao Research on Content Driven Computing Task Offloading and Resource Allocation Algorithm in Connected Vehicles [D]. Beijing: Beijing University of Posts and Telecommunications, 2021
2. [2] Liu Jianhua, Luo Rongxin, Liu Jiajia, etc NSGA - Ⅱ based task unloading scheme for Internet of Vehicles edge computing [J]. Journal of Xi'an University of Technology, 2023, 39 (4): 557-566
3. [3] Ni Junhong, Zang Yun Multi task offloading algorithm under heterogeneous edge cloud architecture [J]. Journal of Harbin Engineering University, 2024, 45 (4): 800-807
4. [4] Lu Weifeng, Yin Wenxu, Wang Jing, etc Cooperative task unloading for vehicle edge computing [J]. Engineering Science and Technology, 2024, 56 (1): 89-98
5. [5] Lin Xinyu, Yao Zewei, Hu Shengxi, etc Task offloading algorithm based on federated deep reinforcement learning in the Internet of Vehicles [J]. Computer Science, 2023, 50 (9): 347-356
6. [6] Deng Wengui, Wen Zhicheng, Ou Jing Edge cloud collaborative task offloading strategy based on deep reinforcement learning [J]. Computer Engineering and Design, 2024, 45 (2): 321-327
7. [7] Zhang Haibo, Liu Xiangyu, Jing Kunlun, etc Research on Unloading Strategy Based on NOMA-MEC in Connected Vehicles [J]. Journal of Electronics and Information Technology, 2021, 43 (4): 1072-1079
8. [8] Pan Shuai Research on task unloading algorithm of mobile edge computing enabled Internet of Vehicles [D]. Nanjing: Nanjing University of Posts and Telecommunications, 2022
9. [9] Zhao Haitao, Zhang Tangwei, Chen Yue, etc Task distribution and offloading algorithm for vehicular edge network based on DQN [J]. Journal of Communications, 2020, 41 (10): 172-178
10. [10] Xu Xiaolong, Fang Zijie, Qi Lianyong, etc Distributed service unloading method based on deep reinforcement learning in the Internet of Vehicles edge computing environment [J]. Journal of Computer Science, 2021, 44 (12): 2382-2405

About the authors

Ge Gaoteng (July 1999-), male, Han ethnicity, from Cangzhou, Hebei, master's student, research interests include task offloading and resource allocation in vehicular edge computing.