Multi-User Competitive Offloading Strategy Based On Mobile Prediction On MEC

With the development of network and communication technology,the explosive popularity of mobile user equipment(UE)has accelerated the emergence and development of many delay-sensitive and computationally intensive applications and services.The computing,storage and network resources of UE Management optimization is facing new problems.As a result,the mobile edge computing(MEC)solution came into being.MEC refers to migrating tasks on the UE to the MEC server to reduce the resource pressure of the UE,but UE mobility sometimes make tasks offloading produce unexpected energy consumption,and at the same time,the competition between multiple UEs for the resources will also cause congestion,which greatly limits the improvement of MEC performance.Therefore,adding mobile prediction technology to the task offload strategy while considering a more efficient the strategy of offloading is an effective means to reduce system energy consumption,achieve green computing,and improve service quality.This thesis mainly studies the energy consumption optimization strategy when multiple mobile user equipments in the MEC system compete for limited resources.The main achievements include:1.To solve the problem that the user equipment in the MEC system produce unexpected energy consumption because of the changes of position,this thesis apply mobile prediction technology to the strategy of offloading and predict the next location according to the user's historical stay area sequence by the algorithm,Prediction by Partial Matching with Dynamic Parameter Updates(PPMDP).The model realizes the prediction of the next position through a generalized hybrid mechanism and a dynamic parameter update mechanism.The experiments show that the strategy in this thesis e has better location prediction accuracy and effectively reduces the expected external energy consumption of the MEC system..2.Aiming at the unexpected energy consumption in the process of multi-task competitive offloading,this thesis apply a finer-grained priority-based mutation mechanism to the Firefly Algorithm(FA),and proposes a strategy of task offloading based on Mutated Firefly Algorithm(MFA).The strategy calculates the data transmission rate based on the prediction results of the first part,the data transmission rate results,and combines relevant parameters to establish an energy consumption optimization model.Then,the improved algorithm is applied to solve the optimization problem.Through continuous iteration of the mutant firefly algorithm,the optimal user equipment in the MEC system is obtained Unloading decision.The results of experiments show that the multi-task competitive offloading strategy based on mutant firefly proposed in this thesis effectively reduces the energy consumption of the MEC system compared with other optimization strategies.