Research On Offloading Strategy Of Multi-User Cooperative Mobile Edge Computing Based On D2D

The rapid development of the Internet of Things has spawned a large number of computing-intensive applications,which has brought great challenges to the data processing of mobile communication networks.Due to the limitations of the mobile device,it is difficult to meet the ultra-low latency requirements of emerging applications.Mobile Edge Computing(MEC)puts the server on the side close to the user,and uses the computing resources of the server to improve the computing power of the device.As the network load further increases,the pressure on the MEC server increases accordingly.At this time,a device-to-device(D2D)communication technology can be introduced.Through the computing and storage capabilities of the mobile terminal equipment,it cooperates with the traditional MEC infrastructure to provide services for other mobile terminal equipment.Based on the scenario of the D2D-assisted MEC computing,current research at home and abroad mainly focuses on the selection of node offloading strategies and the allocation of resources,but the cooperation between nodes is often ignored.Therefore,it is necessary to design an incentive measure to promote collaboration among users in order to better relieve the pressure on MEC servers.In addition,the scenarios of user collaboration in the current research are often static.Considering that the nodes are moving in the actual scenario,how to ensure the stability of data transmission and solve the problem of mobile perception in dynamic scenarios is a major challenge.Based on the above analysis,the main work of this paper is as follows:First,a static scenario based on D2D-assisted offloading is studied under the coverage scenario of a single MEC server.In this scenario,the offloading strategy and resource allocation scheme are jointly optimized to minimize the total system delay.In order to achieve this goal,after analysis,it is necessary to promote cooperation among users,so the bilateral matching theory is introduced,and a preference set based on comprehensive performance indicators and comprehensive execution efficiency is designed.The preference set enables computing nodes and resource nodes to form a stable alliance,which effectively motivates users to perform offloading operations.The simulation results show that the algorithm can effectively reduce the system delay and improve the overall performance of the system.Secondly,based on the above research,in the scenario where the node is in a moving state,the system delay and energy are comprehensively considered,and a system comprehensive utility function is proposed,which can reflect the performance improvement of offload computing compared to the local computing.At the same time,in order to ensure the stability of data transmission during the unloading process in dynamic scenarios and the long-term benefits of the system,a multi-node cooperative scheduling algorithm based on the deep reinforcement learning algorithm is proposed.After simulation verification,the algorithm can effectively motivate users to perform unloading operations and ensure the stability of data transmission,so that the system can obtain higher long-term benefits.