{"title":"基于mec的车载网络移动感知QoS提升与负载均衡:一种深度学习方法","authors":"Chih-Ho Hsu, Yao Chiang, Yi Zhang, Hung-Yu Wei","doi":"10.1109/VTC2021-Spring51267.2021.9448705","DOIUrl":null,"url":null,"abstract":"Recently, Multi-access Edge Computing (MEC) has become a promising enabler to support emerging applications in vehicular networks by offloading compute-intensive tasks from vehicles to proximate MEC servers. However, the high mobility of vehicles brings difficulties to provide reliable services in the MEC system due to potential outages of communication in the process of offloading. Also, load balancing of the MEC system is seldom considered in previous offloading schemes, which may increase the risk of system failure and reduce Quality of Service (QoS) of vehicles due to congestions. Currently, we still lack a low-complexity method to address these issues. In this paper, we aim to promote QoS of vehicular applications by taking vehicles' mobility and latency requirements into account while guaranteeing load balancing of the MEC system. Specifically, we first formulate the joint offloading decision and resource allocation problem as a Mixed Integer NonLinear Programming (MINLP) problem. Then, by taking advantage of both Deep Neural Network (DNN) and Particle Swarm Optimization (PSO), we propose a novel framework to effectively address the problem, where PSO accelerates the training by providing high quality labeled data to DNN. Finally, simulation results show that our proposed method outperforms traditional heuristic algorithms in terms of QoS and runtime.","PeriodicalId":194840,"journal":{"name":"2021 IEEE 93rd Vehicular Technology Conference (VTC2021-Spring)","volume":"66 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2021-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"7","resultStr":"{\"title\":\"Mobility-Aware QoS Promotion and Load Balancing in MEC-Based Vehicular Networks: A Deep Learning Approach\",\"authors\":\"Chih-Ho Hsu, Yao Chiang, Yi Zhang, Hung-Yu Wei\",\"doi\":\"10.1109/VTC2021-Spring51267.2021.9448705\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Recently, Multi-access Edge Computing (MEC) has become a promising enabler to support emerging applications in vehicular networks by offloading compute-intensive tasks from vehicles to proximate MEC servers. However, the high mobility of vehicles brings difficulties to provide reliable services in the MEC system due to potential outages of communication in the process of offloading. Also, load balancing of the MEC system is seldom considered in previous offloading schemes, which may increase the risk of system failure and reduce Quality of Service (QoS) of vehicles due to congestions. Currently, we still lack a low-complexity method to address these issues. In this paper, we aim to promote QoS of vehicular applications by taking vehicles' mobility and latency requirements into account while guaranteeing load balancing of the MEC system. Specifically, we first formulate the joint offloading decision and resource allocation problem as a Mixed Integer NonLinear Programming (MINLP) problem. Then, by taking advantage of both Deep Neural Network (DNN) and Particle Swarm Optimization (PSO), we propose a novel framework to effectively address the problem, where PSO accelerates the training by providing high quality labeled data to DNN. Finally, simulation results show that our proposed method outperforms traditional heuristic algorithms in terms of QoS and runtime.\",\"PeriodicalId\":194840,\"journal\":{\"name\":\"2021 IEEE 93rd Vehicular Technology Conference (VTC2021-Spring)\",\"volume\":\"66 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2021-04-01\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"7\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2021 IEEE 93rd Vehicular Technology Conference (VTC2021-Spring)\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/VTC2021-Spring51267.2021.9448705\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2021 IEEE 93rd Vehicular Technology Conference (VTC2021-Spring)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/VTC2021-Spring51267.2021.9448705","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Mobility-Aware QoS Promotion and Load Balancing in MEC-Based Vehicular Networks: A Deep Learning Approach
Recently, Multi-access Edge Computing (MEC) has become a promising enabler to support emerging applications in vehicular networks by offloading compute-intensive tasks from vehicles to proximate MEC servers. However, the high mobility of vehicles brings difficulties to provide reliable services in the MEC system due to potential outages of communication in the process of offloading. Also, load balancing of the MEC system is seldom considered in previous offloading schemes, which may increase the risk of system failure and reduce Quality of Service (QoS) of vehicles due to congestions. Currently, we still lack a low-complexity method to address these issues. In this paper, we aim to promote QoS of vehicular applications by taking vehicles' mobility and latency requirements into account while guaranteeing load balancing of the MEC system. Specifically, we first formulate the joint offloading decision and resource allocation problem as a Mixed Integer NonLinear Programming (MINLP) problem. Then, by taking advantage of both Deep Neural Network (DNN) and Particle Swarm Optimization (PSO), we propose a novel framework to effectively address the problem, where PSO accelerates the training by providing high quality labeled data to DNN. Finally, simulation results show that our proposed method outperforms traditional heuristic algorithms in terms of QoS and runtime.