Joint Optimization of Path Planning and Resource Allocation in Mobile Edge Computing
Yu Liu,Yong Li,Yong Niu,Depeng Jin
DOI: https://doi.org/10.1109/tmc.2019.2922316
IF: 6.075
2020-09-01
IEEE Transactions on Mobile Computing
Abstract:With the rapid development of mobile applications, mobile edge computing (MEC), which provides various cloud resources (e.g., computation and storage resources) closer to mobile and IoT devices for computation offloading, has been broadly studied in both academia and industry. However, due to the limited coverage of static edge servers, the traditional MEC technology performs badly in a nowadays environment. To adapt the diverse demands, in this paper, we propose a novel mobile edge mechanism with a vehicle-mounted edge (V-edge) deployed. Aiming at maximizing completed tasks of V-edge with sensitive deadline, the problem of joint path planning and resource allocation is formulated into a mixed integer nonlinear program (MINLP). By utilizing the piecewise linear approximation and linear relaxation, we transform the MINLP into a mixed integer linear program (MILP). To obtain the near-optimal solution, we further develop a gap-adjusted branch & bound algorithm, also called GA-B&B algorithm. Moreover, we propose a low-complexity $L$<math>L</math>-step lookahead branch scheme (referred to as $L$<math>L</math>-step scheme) for efficient scheduling in large-scale scenarios. Extensive evaluations demonstrate the superior performance of the proposed scheme compared with the traditional static edge mechanism. Furthermore, the proposed $L$<math>L</math>-step scheme achieves close performance to the near-optimal solution, an- significantly improves the task completion percentage of state-of-the-art schemes by over 10 percent.
computer science, information systems,telecommunications