The characteristics of UAV-assisted edge computing-low cost, high mobility, and fast response speed-render it more suitable for the marine environment. Deploying edge servers in UAVs is a feasible approach to enhance the performance of the marine communication network; however, resource and energy constraints exist when applying UAVs to the marine network. This paper is based on the UAV-assisted edge computing architecture with NOMA. We establish multi-dimensional variable coupling constraints for task completion timeframe, task offloading decision, power, computational resources, and UAV trajectory. We design a weighted minimization model for system energy consumption. We propose a vertically layered alternating iteration optimization scheme. To escape local optimal solutions effectively, a simulated annealing algorithm is employed for the integer planning problem in the outer layer. The inner layer problem is decomposed into subproblems: user device transmit power, computational resource allocation, and corresponding UAV trajectory planning, given a task offloading decision scheme. However, these decomposed subproblems remain non-convex. To address this, the constraint structure is converted, and the problem is approximated and solved as a convex optimization problem using successive convex approximations. The system utility optimization is approximated using a two-layer algorithm and iterative optimization of subproblems. Simulation experiments show that the proposed joint optimization scheme not only reduces the total energy consumption of the system but also outperforms the scheme under OMA in optimizing task offloading costs.
Energy-Oriented Offloading Decision and Multidimensional Resource Allocation for UAV-Assisted Edge Computing Systems
24.06.2024
1856828 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Resource Allocation and Offloading Strategy for UAV-Assisted LEO Satellite Edge Computing
DOAJ | 2023
|DOAJ | 2024
|