UAV-Assisted Vehicular Edge Computing System: Min-Max Fair Offloading and Position Optimization

The rapid development of edge computing provides low latency for computation-intensive application in Internet of Vehicles. The unmanned aerial vehicles (UAVs) with computation capacity have been deployed to provide both relay and execution service in vehicular edge computing (VEC) system. In this paper, we investigate a heterogeneous UAV-assisted VEC system with multiple vehicular user devices (VUDs) and edge servers (ESs). We model our optimization goal as a problem of minimizing the maximum cost of system in terms of weighted sum of delay and energy consumption. As the research question involves binary optimization variables, UAV position variables, and the form of minimizing the maximum value, it is a mixed-integer nonlinear programming (MINLP) problem, making the solution of this question challenging. We propose an iterative CORAP algorithm that jointly optimizes the Computation Offloading decision, Resource Allocation, alongside the Positioning of the UAV, in order to obtain a feasible solution to the problem. Particular, the offloading decision is transformed into a quadratically constrained quadratic program (QCQP) formulation and obtained through the semidefinite relaxation (SDR) method, the computation and communiation resources allocation decision is obtained by the bisection method, and the optimal position of UAV is derived by the successive convex approximation (SCA) method. Finally, simulation experiments are conducted to evidence the effectiveness and feasibility of our proposed algorithm.