Optimization of Vehicular Edge Computing Under Time-Varying Fading Channels With Path Prediction

This paper studies the design of vehicular edge computing networks (VECNs) with multiple moving vehicles and roadside units (RSUs). Uniquely, our study reflects a pragmatic situation where wireless channels are time-varying in the duration of task offloading and vehicles can travel with inconstant s...

Full description

Saved in:
Bibliographic Details
Published in:IEEE internet of things journal 2024-10, p.1-1
Main Authors: Hu, Dieli, Yuan, Mingang, Huang, Gaofei, Zhao, Sai, Tang, Dong
Format: Article
Language:English
Subjects:
Costs
Delays
Fading channels
Internet of Things
mobile edge computing
Optimization
path prediction
Resource management
Servers
task offloading
time-varying channels
Vehicle-to-infrastructure
vehicular network
Wireless communication
Wireless sensor networks
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper studies the design of vehicular edge computing networks (VECNs) with multiple moving vehicles and roadside units (RSUs). Uniquely, our study reflects a pragmatic situation where wireless channels are time-varying in the duration of task offloading and vehicles can travel with inconstant speeds in a real-world scenario. By jointly optimizing transmit power and time allocation for task offloading as well as computation task partition in the VECN, our goal is to minimize the cost at the vehicles for energy consumption on task offloading and computing, and rent on task computing service at RSUs. However, solving the formulated optimization problem directly is impossible due to the requirement of noncausal vehicular position information (VPI) and noncausal channel state information (CSI) between vehicles and RSUs. To address this issue, a path prediction model is adopted to predict the noncausal VPI, based on which the noncausal CSI can be estimated. Then, a novel receding horizon optimization method is proposed to transform the original problem into a sequence of tractable problems. Despite this, the problems remain complex due to the computationally prohibitive task of identifying the optimal task offloading duration at each vehicle in a centralized manner. To overcome this difficulty, the consensus alternating directions method of multipliers is proposed to solve the problem in a distributed manner with low computational complexity. Numerical results show that our proposed scheme can save at most 30% of monetary cost as compared with existing baseline schemes.
ISSN:2327-4662
2327-4662
DOI:10.1109/JIOT.2024.3488193