Vehicle Speed Aware Computing Task Offloading and Resource Allocation Based on Multi-Agent Reinforcement Learning in a Vehicular Edge Computing Network

TL;DR

This paper introduces a vehicle speed aware strategy for task offloading and resource allocation in vehicular edge computing, optimizing energy use while respecting delay constraints based on vehicle speed and task type.

Contribution

It proposes a novel vehicle speed aware delay model and employs multi-agent reinforcement learning to optimize task offloading and resource allocation.

Findings

Reduces energy cost of task execution.

Maintains delay constraints across different vehicle speeds.

Outperforms existing strategies in simulations.

Abstract

For in-vehicle application, the vehicles with different speeds have different delay requirements. However, vehicle speeds have not been extensively explored, which may cause mismatching between vehicle speed and its allocated computation and wireless resource. In this paper, we propose a vehicle speed aware task offloading and resource allocation strategy, to decrease the energy cost of executing tasks without exceeding the delay constraint. First, we establish the vehicle speed aware delay constraint model based on different speeds and task types. Then, the delay and energy cost of task execution in VEC server and local terminal are calculated. Next, we formulate a joint optimization of task offloading and resource allocation to minimize vehicles' energy cost subject to delay constraints. MADDPG method is employed to obtain offloading and resource allocation strategy. Simulation results show that our algorithm can achieve superior performance on energy cost and task completion delay.