This paper addresses the problem of long-term service inequality of IoT devices in UAV-assisted wireless power communication networks. Considering factors such as the number of device historical services, the amount of device data, the UAV flight trajectory, and the binary service decision of the device, the scheduling and trajectory planning for UAV-assisted wireless power supply network data collection are modeled as a multi-objective optimization problem. The goals are to minimize the UAV’s energy usage, maximize the overall volume of data gathered, and guarantee long-term equity in device maintenance. Given the NP-hardness and mixed integer non-linear nature of the problem, a multi-objective master–slave co-optimization (MOMSCO) algorithm is proposed. The algorithm breaks the optimization problem down into a primary module and smaller modules for the solution. The primary module designs a multi-objective dynamically balanced sequence generating algorithm, chooses the best device service sequence, and calls the sub-module to optimize the UAV’s flying trajectory. By building a lightweight state space and creating a multi-objective reward function, the submodule proposes an expanded Deep Deterministic Policy Gradient (DDPG) algorithm to co-optimize UAV energy consumption and data gathering. Numerical simulation is used to validate the real scenario, and the findings demonstrate that the device service’s fairness significantly improves, steadily converging to about 97% after several data collection cycles. The usefulness of the suggested algorithm in enhancing system performance and long-term service fairness is further demonstrated by the fact that the system energy efficiency of the proposed scheme beats that of the benchmark scheme in every round.
扫码关注我们
求助内容:
应助结果提醒方式:
