Wireless Networks ( IF 3 ) Pub Date : 2019-09-05 , DOI: 10.1007/s11276-019-02127-y Xiaolong Xu , Renhao Gu , Fei Dai , Lianyong Qi , Shaohua Wan
Abstract
The Internet of Vehicles (IoV) has gained worldwide attentions as it provides the service of collecting real-time traffic information to improve the road safety. The IoV users can offload their computing tasks to the edge computing devices (ECDs) for low latency execution and the cloud can be engaged to process big data with sufficient computing resources. Though galactic convenience brought by the IoV cloud-edge computing system, it remains a challenge to manage the resource of ECDs by reducing the energy and time consumption while avoiding the situation of overload or underload of the ECDs to maintain the system-stability. Moreover, during the movement of the vehicles, the computing tasks and data may be uploaded to different ECDs and the data continuity may be destroyed. In this paper, a multi-objective computation offloading method (MOC) for IoV in cloud-edge computing is proposed to deal with the challenges above. A vehicle-to-vehicle communication-based route obtaining algorithm is designed first. Then, in order to ensure the trustworth of the IoV data, which ECD to upload the computing tasks to is selected. Under the case that all ECDs are overloaded, the computation offloading between ECDs and cloud is considered. In addition, non-dominated sorting genetic algorithm III is adopted to realize the multi-objective optimization of decreasing the load balancing rate and reduce the energy consumption in ECDs and shorten the time during processing the computing tasks. Furthermore, we employ the simple additive weighting and multiple criteria decision making to evaluate the solutions of our proposed method. Finally, experimental evaluations are conducted to validate the efficiency and effectiveness of our proposed method.
中文翻译:
云边缘计算中的车联网多目标计算分流
摘要
车辆互联网(IoV)提供了收集实时交通信息以改善道路安全的服务,因此受到了全世界的关注。IoV用户可以将其计算任务转移到边缘计算设备(ECD)上以实现低延迟执行,并且可以使用云来处理具有足够计算资源的大数据。尽管IoV云边缘计算系统为银河带来了便利,但通过减少能耗和时间消耗,同时避免ECD的过载或欠载情况来维持系统稳定性,管理ECD的资源仍然是一个挑战。此外,在车辆运动期间,计算任务和数据可以被上载到不同的ECD,并且数据的连续性可能被破坏。在本文中,针对上述挑战,提出了一种用于云边缘计算中IoV的多目标计算卸载方法(MOC)。首先设计了基于车载通信的路径获取算法。然后,为了确保IoV数据的可信度,选择将计算任务上载到哪个ECD。在所有ECD都过载的情况下,考虑了ECD与云之间的计算分流。另外,采用非支配排序遗传算法III,实现了降低负载均衡率,降低ECD能耗,缩短计算任务处理时间的多目标优化。此外,我们采用简单的加法加权和多准则决策来评估我们提出的方法的解决方案。最后,