Optimisation Algorithm for Offloading Decision and Resource Allocation in Cache‐Assisted Multi‐UAV Mobile Edge Computing Systems

ABSTRACT

With the booming development of Internet of Things applications, mobile edge computing (MEC) has become a key technology to handle massive data processing demands. Unmanned aerial vehicles (UAVs), with their advantages of high mobility and flexible deployment, are becoming important nodes in constructing an integrated air‐ground MEC network. Targeting the cache‐assisted UAV MEC scenario, this paper proposes a joint optimisation algorithm for UAV positioning, offloading decisions and resource allocation. This algorithm allows for the processing of computational tasks across time slots and considers the immediate reuse of computational resources upon release. Under the constraints of UAV cache capacity, heterogeneous computing capabilities and communication bandwidth, an optimisation problem is formulated to minimise the weighted sum of the average task completion time and user costs. First, addressing the user distribution characteristics, this paper uses a location‐based clustering algorithm to determine the initial deployment of UAVs and optimises their hovering positions through a local search strategy to maximise user coverage. For dynamically arriving computational tasks, this paper proposes an offloading decision and resource allocation optimisation algorithm based on a genetic algorithm‐particle swarm optimisation (GA‐PSO) approach. The original problem is decomposed into two subproblems: offloading decision and resource allocation. Comprehensively considering communication and computation performance, a fitness function integrated with a discrete‐event simulator is designed. A genetic algorithm and a particle swarm optimisation algorithm are used to solve these respective subproblems. Simulation results show that the proposed GA‐PSO algorithm can effectively reduce task completion time and user costs.