Research On Optimization Algorithm Of Energy Management For Multi-user Mobileedge Computing System

With the rapid development of mobile Internet,people have been used to arranging their daily life through a variety of mobile terminals,but at present,the limitations of mobile devices in computing capacity,storage capacity and battery capacity make it difficult for users to fully meet their computing needs in practice.How to effectively improve the contradiction between the limited resources of mobile devices and the needs of users is undoubtedly a great challenge.As a new computing framework,Mobile Edge Computing(MEC)transmits some computing intensive data tasks to MEC server for calculation by making effective unloading decision,so as to relieve the computing pressure on terminal equipments,reduce system energy consumption,ensure the quality of computing service,and provide users with a high bandwidth and low delay computing service environment.Aiming at the problem of energy management in MEC system,this paper studies the problem of system energy consumption optimization in the scenario of multi-user single MEC server equipped with energy harvesting equipment,and proposes an optimization algorithm based on Lyapunov method and a distributed optimization algorithm based on the Alternative Direction Method of Multipliers(ADMM),which realize the optimization goal of minimizing system energy consumption under the condition of satisfying certain quality of service and system queue stability.The main work and innovations are summarized as follows:First,the MEC system considered in this paper includes a number of mobile terminals equipped with energy harvesting devices and an MEC server.The data tasks generated by the applications running on the terminal can be assigned to local computing or transmitted to MEC server for computing.Under the constraints of ensuring the quality of service and the stability of battery queue,this paper constructs a system energy consumption minimization problem.In order to solve the problems that the randomness of the collected renewable energy and the task coming,and the time coupling of the queue,an optimization algorithm based on the Lyapunov method is proposed.By using the Lyapunov drift plus penalty function,the original problem is transformed into solving the system queue stability problem and minimizing the system energy consumption meanwhile.The asymptotic optimality and the stability of the system queue are proved by a series of theoretical analysis.Secondly,considering the problem that the computing complexity increases with the largescale access request brought by the increase of users,and the data privacy security of users.Based on the Lyapunov optimization algorithm,this paper proposes a distributed optimization algorithm based on ADMM method.By means of decomposition and coordination,the original problem is decomposed into three subproblems,and the optimal solution of the original problem is obtained by coordinating and optimizing each local subproblem through iterative updating.Under the distributed algorithm,each user first optimizes one subproblem independently,and then sends the result to the server for subsequent processing.Therefore,the user does not need to provide all the information to the MEC server,so as to achieve the purpose of protecting the user's data privacy.Finally,by selecting the appropriate experimental data and system parameters for MATLAB simulation,the experimental results show that the average energy consumption of the system is significantly reduced under the algorithm,and the battery queue stability and quality of service constraints are always met under all time slots,which proves the effectiveness and feasibility of the proposed algorithm.