Research On Offloading Strategy Of Energy Harvesting Technology Combined With Edge Computing Systems

With the rapid development of wireless communication technology,wireless devices are facing more complex and changeable scenarios than ever before.The traditional network architecture has been unable to meet the needs of delay sensitive applications.Mobile Edge Computing(MEC)technology transfers the computing and storage capacity to the network edge,which solves the delay problem of traditional network architecture.At the same time,wireless devices can not meet the delay requirements of intensive computing tasks due to the limitation of their battery capacity,and Energy Harvesting(EH)technology can realize wireless charging of the battery of wireless devices and enhance the operation ability of wireless devices.Based on the idea of MEC technology,academia has proposed a wireless device with EH device to assist in task computing.By unloading the computing task of the wireless device to the edge server for computing,users can experience low latency services.In addition,since wireless devices can be powered by free solar or wind energy provided by EH technology,energy consumption is not an urgent problem.At present,most of the relevant research work focuses on the unloading strategy and resource allocation in the unloading process of MEC technology,while the research on MEC system with EH equipment is relatively rare.Focusing on the delay sensitivity and computational efficiency of wireless devices,this thesis makes a detailed study on the unloading matching strategy of MEC system with energy acquisition devices and multiple edge servers.The main contributions are as follows:(1)Aiming at the MEC application scenario with EH equipment,the EH technology and MEC technology in this scenario are introduced respectively.Combined with the characteristics of the two technologies,the MEC system architecture with EH equipment is analyzed.Then,the classical optimization algorithm for solving the system objective problem is introduced in detail and the complexity is analyzed.(2)Aiming at the problem of computing resource management in the network,a binary unloading MEC system model based on EH is constructed.The model adopts binary unloading strategy,considers the loss caused by over charging and discharging of wireless device battery,as well as the energy constraints of wireless device and the time variability of wireless channel,and aims to maximize the data computing rate of wireless device and the proportion of computing tasks successfully executed to the total tasks.To solve the objective problem,a scheme of jointly optimizing task unloading decision and wireless resource allocation under binary unloading strategy is proposed.Simulation results show that the proposed strategy is suitable for time-varying channel conditions.On the basis of ensuring the battery life,it maximizes the data computing rate of wireless devices and the proportion of computing tasks successfully executed to the total tasks.(3)Aiming at the problems of wireless device delay and device mobility,a multi-user and multi server MEC system model based on EH is constructed.According to the characteristics of multi MEC server and multi-user,based on the above single MEC server model,this model considers the unloading matching between multi MEC server and multi-user,as well as the mobility of wireless devices.To solve this model,a lyapunov optimal resource allocation algorithm based on predicting equipment trajectory is proposed.The algorithm does not need to quantify the system state and feasible action set,and the algorithm decision complexity in each time slot is low.Considering the mobility of wireless devices,through the Viterbi optimization algorithm based on Hidden Markov Model(HMM),the optimal unloading MEC server under the real-time movement of wireless devices is solved,which reduces the switching times of MEC server unloading,achieves better performance than the single MEC computing system,and further reduces the delay of wireless devices.