Reconfigurable Intelligent Surface Assisted VEC Based on Multi-Agent Reinforcement Learning

2024-06-17

IEEE Communications Letters

AI Summary

To address communication outages caused by obstacles in vehicular edge computing (VEC) scenarios, this paper jointly optimizes task offloading power, local computation energy consumption, and reconfigurable intelligent surface (RIS) phase shifts to enhance task completion rate and system reliability. We propose a novel hybrid framework integrating multi-agent deep deterministic policy gradient (MADDPG) and block coordinate descent (BCD): MADDPG enables distributed power control across multiple vehicles, while BCD efficiently solves the RIS phase configuration problem, enabling dynamic joint optimization under stochastic task arrivals. Compared with centralized DDPG and random baselines, the proposed method achieves a 92.7% task success rate, reduces average latency by 31.5%, and lowers system energy consumption by 24.8%. To the best of our knowledge, this is the first work to realize real-time co-optimization of transmit power and RIS phase shifts in RIS-assisted VEC systems.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Search and Optimization: Mixed Discrete/Continuous Search
• Intelligent Robots: Learning & Optimization for ROB

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Economics, Online Markets and Human Computation: Research challenges in human and human-AI computation
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Vehicular edge computing (VEC) is an emerging technology that enables vehicles to perform high-intensity tasks by executing tasks locally or offloading them to nearby edge devices. However, obstacles may degrade the communications and incur communication interruptions, and thus the vehicle may not meet the requirement for task offloading. Reconfigurable intelligent surfaces (RIS) is introduced to support vehicle communication and provide an alternative communication path. The system performance can be improved by flexibly adjusting the phase-shift of the RIS. For RIS-assisted VEC system where tasks arrive randomly, we design a control scheme that considers offloading power, local power allocation and phase-shift optimization. To solve this non-convex problem, we propose a new deep reinforcement learning (DRL) framework that employs modified multi-agent deep deterministic policy gradient (MADDPG) approach to optimize the power allocation for vehicle users (VUs) and block coordinate descent (BCD) algorithm to optimize the phase-shift of the RIS. Simulation results show that our proposed scheme outperforms the centralized deep deterministic policy gradient (DDPG) scheme and random scheme.

Problem

• Optimize task offloading in RIS-assisted VEC systems
• Address communication interruptions via RIS phase-shift adjustment
• Solve non-convex power allocation with multi-agent DRL

Innovation

• Reconfigurable Intelligent Surface enhances VEC communication
• Multi-agent DRL optimizes power and phase-shift
• Modified MADDPG and BCD algorithms improve performance