An RSU-crossed dependent task offloading scheme for vehicular edge computing based on deep reinforcement learning. (27th April 2023)

• Record Type: Journal Article
• Authors: Bi, Xiang; Shi, Jianing; Zhang, Benhong; Lyu, Zengwei; Huang, Lingjie
• Is Part Of: International journal of sensor networks. Volume 41:Number 4(2023)
• Page Start: 244
• Page End: 256
• Publication Date: 2023-04-27
• Subjects: task offloading -- vehicular edge computing -- VEC -- dependent task -- deep reinforcement learning -- DRL; Sensor networks -- Periodicals; 681.2
• Languages: English
• ISSNs: 1748-1279

Abstract

Various interdependent and computationally intensive on-vehicle tasks have posed great pressure on the computing power of vehicles. Vehicular edge computing (VEC) is considered to be a promising paradigm to solve this problem. However, due to the high mobility, vehicles will pass through multiple road-side units (RSUs) during task computing. How to coordinate the offloading decision of RSUs is a challenge. In this study, we propose a dependent task offloading scheme by considering vehicle mobility, service availability, and task priority. Meanwhile, to coordinate the offloading decisions among the RSUs, a Markov decision process (MDP) is carefully designed, in which the action of each RSU is divided into three steps to decide whether, where, and how each task is offloaded separately. Then, an advanced DDPG-based deep reinforcement learning (DRL) algorithm is adopted to solve this problem. Simulation results show that the proposed scheme has better performance in reducing task processing latency and consumption.