Research On Energy Consumption Optimization In Mobile Edge Computing Networks

In recent years,mobile edge computing technology has gotten widespread industry attention because of its high bandwidth and low latency user experience.This technology shortens the transmission distance by densely deploying base stations and edge servers near the users,which achieves the purpose of increasing the transmission rate and reducing the delay.However,since the number of users and the amount of communication fluctuate over time,it will cause a great waste for those edge servers whose resources are underutilized.On one hand,due to the unreasonable scheduling of service requests to users,the utilization of edge servers is low.On the other hand,catering to user demand during peak traffic,the data dispatching center usually opens the corresponding edge servers according to the demand of traffic peaks.It may cause a large number of edge servers to be idle when traffic is low.Therefore,in the face of the current trend of deploying a large number of edge base stations,opening an appropriate number of edge servers at different times is an effective way to save energy for mobile edge computing networks.Based on the existing mobile edge computing network,this thesis establishes an energy consumption model,sets the optimization objectives and related constraints,and proposes corresponding network energy optimization algorithms for the mobile edge computing networks including delay constraints and delay-free constraints,respectively.Aiming at the network energy optimization problem without delay constraints in the mobile edge computing architecture,in this thesis,it is defined as the network energy minimization problem and a network energy optimization algorithm is designed.Since the number of users and traffic in the network fluctuates with time,it is assumed that each edge base station can serve users within its connection range and can meet its computing and transmission requirements.This thesis simulates the curve of the number of users with time,and proposes a lightweight linear regression prediction mechanism to predict the user service requests in the next time slice.Each station makes utilization prediction of the edge server according to the number of requests for each service of the next time slice and then make relevant decisions.In this problem,the user connection between the adjacent base stations,service configuration and the conditions of closing the edge server are taken into comprehensive consideration.The policies are designed to migrate the users.The experiment results show that the proposed algorithm can reduce the energyconsumption of the network and reduce the switching energy consumption of the edge servers compared with the existing algorithms.As for the network energy optimization problem with delay constraints in the mobile edge computing architecture,in this thesis,we also propose a heuristic delay-limited network energy optimization algorithm.In some real life scenarios,some tasks need to be completed within the specified time.Therefore,based on the previous problem,the problem is considered comprehensively for the computation delay and transmission delay.Therefore,this problem has become more complicated due to the increased delay constraints.So the problem has been remodeled.In addition,the users connection,the services configuration,the mechanism for selecting the edge server to be closed,and the strategy for migrating users have been changed and improved,so that the network energy consumption can be optimized to the maximum extent while satisfying the user service quality.The experiment results show that the proposed network energy optimization algorithm with time delay constraints can effectively reduce the network energy consumption compared with other existing algorithms.