Lyapunov-Based Joint Flight Trajectory and Computation Offloading Optimization for UAV-Assisted Vehicular Networks

In recent years, UAV-assisted mobile edge computing (MEC) has attracted significant attention. However, it is still challenging to dispatch a UAV to accompany ground vehicles and provide both communication and computation support in a highly dynamic environment with various constraints on mobility, coverage, and resources. This study delves into a novel, low-complexity, long-term UAV-assisted vehicular cooperative computation problem, examining the reciprocal impact of vehicles' flight/driving trajectories and the complementary relationship among different offloading options. Specifically, we formulate a joint optimization problem that considers flying trajectory and offloading decision, aiming to minimize both service delay and energy consumption from a long-term perspective. Due to the time coupling of variables, we employ the Lyapunov optimization framework to decompose the original problem into manageable subproblems for each time slot. Furthermore, we introduce a low-complexity Greedy Bats Algorithm (GBA) to solve the NP-hard two-dimensional generalized assignment problem (TDGAP), optimizing the upper bound of the Lyapunov drift-plus-penalty function to minimize service delay in each time slot. Additionally, we utilize the Successive convex approximation (SCA) algorithm to convert the UAV's trajectory optimization problem into a convex problem for further low-complexity solution. Simulation results demonstrate that our proposed scheme outperforms other comparative algorithms in terms of computation delay, complexity and energy consumption.