Energy Efficient Offloading And Resource Allocation For Multi-Access Edge Computing

With the continuous development of communication technology,the number of smart devices and data traffic will explode.The applications on smart devices(such as AR/VR,etc.)have continuously increased the requirements for data transmission,mobile device computing capabilities,and computing latency.Traditional cloud computing provides users with more computing and storage resources,but it does not solve the problem of computing latency.Multi-access edge computing has become an important solution to this problem.Multi-access edge computing transforms part of the cloud's service capabilities to edge nodes of near users,and providing users with computing,cache,and storage services.Multi-access edge computing is one of the key technologies of 5G.It can analyze and process most of the computing tasks generated by users on the MEC server at the edge nodes,and MEC effectively reducing the time delay in the data transmission process,and alleviating the transmission of the backbone network pressure.In addition,computing power is no longer an obstacle to the development of applications and apps in smart devices.The research objective of this thesis is to minimize the energy and cost of computing tasks during the offloading of computing tasks while meeting the user's requirements for time delay in an environment of limited computing resources and wireless channel resources.The main tasks as follows:In the single MEC server offloading solution,this solution uses the core network as one of the calculation task offloading platforms to solve the problem of fast calculation and analysis for large data volume calculation tasks.The core network has huge computing resources,which can meet the computing requirements of some tasks with large data volume.A priority-based algorithm for joint energy efficiency offloading and resource allocation is proposed.The algorithm proposes the concept of priority for each user.The priority of each user is determined by the computing power and delay requirements of the user.A user with a high priority can preferentially select a wireless channel with a better channel condition for data transmission,and use the remaining power of the user as a selection criterion for the priority load factor.The computing capacity of a cell with a single MEC server is limited and cannot meet the increasing number of users and the delay requirements of big data calculation tasks.Therefore,this thesis extends the single-server offloading mode to multi-server offloading mode,and proposes an offloading algorithm based on FFD(First Descending Adaptive Algorithm in Descending Order)and a layered algorithm.Tasks are sorted in descending order by task size,and then loaded into boxes(MEC server)in order.The algorithm uses a hierarchical principle,that is,we allocate the same amount of wireless channel resources to users in the initial stage.In the case of meeting user delay requirements,the computing resources required for each computing task are determined,and then a competition mechanism is introduced to allocate wireless channel resources to users.Simulation results show that the system energy consumption of this scheme is greatly reduced compared to random allocation,and the use efficiency of the MEC server is improved.