Abstract
Vehicular edge computing (VEC) is an emerging technology with significantpotential in the field of internet of vehicles (IoV), enabling vehicles toperform intensive computational tasks locally or offload them to nearby edgedevices. However, the quality of communication links may be severelydeteriorated due to obstacles such as buildings, impeding the offloadingprocess. To address this challenge, we introduce the use of ReconfigurableIntelligent Surfaces (RIS), which provide alternative communication pathways toassist vehicular communication. By dynamically adjusting the phase-shift of theRIS, the performance of VEC systems can be substantially improved. In thiswork, we consider a RIS-assisted VEC system, and design an optimal scheme forlocal execution power, offloading power, and RIS phase-shift, where random taskarrivals and channel variations are taken into account. To address the scheme,we propose an innovative deep reinforcement learning (DRL) framework thatcombines the Deep Deterministic Policy Gradient (DDPG) algorithm for optimizingRIS phase-shift coefficients and the Multi-Agent Deep Deterministic PolicyGradient (MADDPG) algorithm for optimizing the power allocation of vehicle user(VU). Simulation results show that our proposed scheme outperforms thetraditional centralized DDPG, Twin Delayed Deep Deterministic Policy Gradient(TD3) and some typical stochastic schemes.