Unmanned aerial vehicle (UAV)-assisted multiaccess edge computing (MEC) technology has garnered significant attention and has been successfully implemented in specific scenarios. The optimization of the network energy consumption in the relevant scenarios is essential for the whole system performance due to the constrained energy capacity of UAVs. However, the dynamic changes in MEC network resources make the energy consumption optimization a challenge. In this paper, a multi-UAV-multiuser MEC model is established to assess the system energy consumption, and the optimization problem of multi-UAV cooperation strategies is formulated base on the model. Then, a multi-agent deep deterministic policy gradient (MADDPG) algorithm based on deep reinforcement learning (DRL) is employed to resolve the above optimization problem. Each UAV is equivalent to a single agent that cooperates with other agents to train actors and critic evaluation networks to accomplish collaborative decision-making. In addition, a prioritized experience replay (PER) scheme is used to improve the convergence of the training process. Simulation results show that the impact of changes in different network resources on the network energy consumption by comparing the performance of different algorithms. The findings presented in this paper serve as a valuable reference for future work on system performance optimization, specifically in terms of energy efficiency.

Aiming at the difficulties in modelling, simulation and verification in digital twin workshop, a modelling and online training method for digital twin workshop is proposed. This paper describes a multi-level digital twin aggregate modelling method, including the status attributes, the static performance attributes and the fluctuation performance attributes, and designs a digital twin organisation system, namely, digital twin graph. According to the data demand for digital twin aggregates, a spatio-temporal data model is constructed. The digital twin model training method using truncated normal distribution is presented. Furthermore, a verification method based on real-virtual error for a digital twin model is proposed. The effectiveness of real-time status monitoring, online model training and simulation for production is verified by a case.

This paper studies a rotary-wing unmanned aerial vehicle (UAV)-assisted wireless powered fog computing network, where a UAV serves as both a mobile wireless energy source and a mobile fog server to provide charging and computing services to a group of ground sensors. The UAV first broadcasts radio frequency (RF) signals to charge the sensors, and then the sensors use the harvested energy to complete their tasks by local computing and fog computing, where the partial offloading policy is employed. In order to achieve green communication system design, an optimization problem is formulated to minimize the UAV’s energy consumption by jointly optimizing the UAV’s trajectory, the task offloading allocation and computing resource allocation, where the piecewise nonlinear energy harvesting (EH) model, the charging requirements of sensors and velocity constraints of UAV are taken into account together. To solve the non-convex problem, an iterative method is designed based on the successive convex approximation (SCA) theory. Simulation results show that our presented scheme converges within a few iterations, by using which the UAV’s energy consumption can be greatly reduced compared to benchmark schemes. It is also observed that the harvested energy by the linear EH model have obvious bias compared to that obtained by the nonlinear EH model. And in the case of nonlinear EH model, as circuit saturation threshold increases, the UAV moves closer to the sensors such that their tasks can be completed, which may result in a higher energy consumption of the UAV. Besides, the velocity variation of UAV is relatively small by introducing the velocity constraint, which avoids drastic steering of the UAV in our considered scenario.