Task scheduling of computation-intensive graph jobs in UAV-assisted hybrid vehicular networks

• Published in: Vehicular Communications Vol. 42; p. 100630
• Main Authors: Chai, Jiahui, Meng, Yun, Wang, Wei, Lyu, Yue, Yue, Haochuan, Liu, Xinyi
• Format: Journal Article
• Language: English
• Published: Elsevier Inc 01.08.2023
• Subjects: Average completion delay; Computation-intensive graph jobs; Hybrid vehicular networks; Multiple jobs scheduling; Unmanned aerial vehicle; Vehicular edge computing; Vehicular edge computing; Hybrid vehicular networks; Computation-intensive graph jobs; Average completion delay; Unmanned aerial vehicle; Multiple jobs scheduling
• ISSN: 2214-2096
• DOI: 10.1016/j.vehcom.2023.100630

As a promising new form in mobile edge computing network, Vehicular Edge Computing (VEC) can further improve the Quality of Service (QoS) of users. However, when edge servers are damaged or overloaded, the requirements of users are difficult met by VEC servers owing to incomplete connectivity and time-varying channel conditions in dynamic and resource-limited vehicular networks. In this paper, we introduce Unmanned Aerial Vehicle (UAV) to enhance the connectivity and total resources in the vehicular networks. Added to that, we focus on the problem of computation-intensive graph jobs scheduling, where the information interactions among the traditional vehicles, autonomous vehicles and UAV are integrated. And the above scheduling problem is modeled as a binary nonlinear programming problem aiming at minimizing average completion delay of jobs. Then, we develop an effective multiple jobs scheduling scheme by considering the scheduling order of multiple jobs, the remaining processing delay of servers, and the available resources of vehicular networks, which consists of two stages: (1) the multiple jobs priority scheduling algorithm for optimizing the jobs scheduling order; (2) the task offloading decision algorithm for optimizing average completion delay of jobs. Finally, the effectiveness of proposed scheme is proved by extensive simulation experiments. Compared with the other schemes, the proposed scheme can effectively reduce average completion delay of jobs.