网站首页 | 期刊介绍 | 编委会 | 投稿指南 | 在线订阅 | 联系我们 | 同行评议 | 出版声明 | 征稿English
梁裕丞,曹傧.VANET云环境下基于人工神经网络的车辆任务卸载策略[J].重庆邮电大学学报(自然科学版),2020,32(3):336-344. 本文二维码信息
二维码(扫一下试试看!)
VANET云环境下基于人工神经网络的车辆任务卸载策略
Artificial neural network methods for task offloading in VANET cloud
投稿时间:2018-12-20  修订日期:2020-05-30
DOI: 10.3979/j.issn.1673-825X.2020.03.002
中文关键词:  VANET云  任务卸载  神经网络  买卖博弈
English Keywords:VANET cloud  task offloading  neural network  buyer/seller game
基金项目:
作者单位E-mail
梁裕丞 重庆邮电大学 移动通信技术重庆市重点实验室,重庆 400065 15708431962@163.com 
曹傧 北京邮电大学 网络与交换技术国家重点实验室,北京 100876 caobin65@163.com 
摘要点击次数: 115
全文下载次数: 73
中文摘要:
      在VANET(vehicular ad-hoc network)云环境中,由于车辆自身资源受限等原因需要将部分计算密集型任务卸载至周围车辆协同处理。而车辆移动的随机性是影响车辆任务卸载性能好坏的重要因素之一,对此,提出了一种车辆间的卸载任务分配策略。考虑到车辆之间连接时间的随机性,提出一种基于人工神经网络的连接时间预测方法,该方法能够通过对历史数据的学习,较为准确地对未来车辆行驶轨迹进行预测。此外,车辆将空闲资源进行共享意味着自身能耗增加,由于车辆本身的自私性使得车辆不会无偿为周围车辆提供服务。为了激励车辆之间进行协作,制定了一种分布式买卖博弈方法达到车辆资源需求与收益之间的平衡,还设计了一种集中式任务分配策略以获得任务卸载的最大效用。仿真显示,提出的方法在最大化卸载效用与提高任务卸载成功率方面都有较好的性能。
English Summary:
      In the VANET cloud, some computationally intensive tasks need to be offloaded to the surrounding vehicles for coordinated processing due to the limited resources of the vehicle itself. The randomness of vehicle movement is one of the important factors affecting the offloading performance. For this reason, we propose an offloading task allocation strategy between vehicles. Considering the randomness of connection time between vehicles, we adopt a connection time prediction method based on artificial neural network. This method can accurately predict the future vehicle trajectory by learning historical data. In addition, sharing idle resources to other vehicles means increasing energy consumption. Due to the selfishness of the vehicles, they will not provide services for surrounding vehicles without compensation. In order to encourage the cooperation of available vehicles, we have developed a distributed Buyer/Seller game method to achieve a balance between resource demand and payment. Finally, we design a centralized task allocation strategy to get the maximum utility of task offloading. The simulation results show that the proposed method is feasible in maximizing the offloading utility and improving the success rate of task offloading.
HTML    PDF浏览   查看/发表评论  下载PDF阅读器
版权所有 © 2009 重庆邮电大学期刊社  
地址:重庆市 南岸区 重庆邮电大学 期刊社 邮编:400065
电话:023-62461032 E-mail : journal@cqupt.edu.cn
meinv 海贼王论坛