Research On The Method Of Computation Offloading And Computing Resource Allocation In Mobile Edge Computing

With the rapid growth of mobile applications,the limited computing resources and battery capacity of user equipments(UEs)when facing applications have become a bottleneck restricting the user experience.The traditional method that offloads the UE's task to cloud computing platform will lead to a long latency for UE,which could not improve UE's Qo S in real-time application environment.Mobile Edge Computing(MEC)is proposed to sink cloud computing resources in the traditional core network into wireless access network,which could provide users with computing resources.Mobile edge computing can provide users with a close proximity,low latency,high rate service environment and improve the current situation of computing resource shortage of UEs.In a multi-MEC network environment,how to offload UEs' tasks and allocate server computing resources is an important research direction in MEC system.The main contents of this study include:Consider the offloading problem and computation resources allocation problem in a heterogeneous MEC network,we propose a heuristic algorithm based on simulated annealing(SA)to implement the decision of UE task offloading.In order to minimize the total energy consumption of UEs in network,After the task computation offloading decision is determined by SA algorithm,we present the minimum computing resources scheme for UE.Simulation results show that,compared with the traditional MEC network which only considers the terrestrial base station MEC network or the UAV enabled MEC network,the heterogeneous MEC network which considers the UAV-MEC platform and the MBS-MEC platform at the same time is proposed in this paper.The proposed heterogeneous MEC network and the SA based offloading algorithm can significantly improve the total energy consumption of UEs in the network.In view of the heterogeneity of MEC server platform,we propose a distributed Deferred Acceptance Offloading Algorithm(DAOA)to implement the UE task calculation in the network.For the problem of many to one matching between user and MEC platform in DAOA,we use the cost function composed of energy consumption and delay as the preference of UE to MEC platform.For the preference of MEC platform to UE,it consists of the transmission time of UE service request and the calculation time of MEC platform.Simulation results show that compared with the Immediate Acceptance(IA)algorithm,the DAOA has a significant improvement in reducing the total cost function of users.